1 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 TRF INSIGHTS Journalism in the AI era: Opportunities and challenges in the Global South and emerging economies A report by Damian Radcliffe Published January 2025 2 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Contents Foreword 03 Executive s ummary 05 Introduction 10 Snapshot of survey respondents 12 Chapter 1: How is AI being used in newsrooms? 13 Journalistic functions and use 16 AI tools and their impact 19 Non-users 21 Chapter 2: Key challenges and risks 24 Lack of newsroom policies 25 Absence of leadership and training 26 Reduced transparency and accountability 28 Diminished creativity and the erosion of skills and jobs 30 Perpetuating bias and misinformation 32 Chapter 3: Moving f orward 35 Attitudes towards AI and its impact on the future of journalism 36 The year ahead: a focus on regulation and ethical standards 38 Looking to the future: AI adoption goes increasingly mainstream 42 Conclusion and looking ahead 44 Recommendations for stakeholders 47 Recommendations for journalists 48 Recommendations for editors and newsroom leaders 49 Recommendations for media development organisations and funders 50 Recommendations for policy makers 51 Recommendations for platforms and developers 52 Appendix 1: Methodology and participant sample 53 Appendix 2: Survey questions 56 Appendix 3: Suggested reading & resources 62 Appendix 4: About the author 64 About the Thomson Reuters Foundation 66 3 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Foreword ANTONIO ZAPPULLA CEO, THOMSON REUTERS FOUNDATION It’s hard to believe that we’re still in the early stages of the generative AI revolution, considering ChatGPT’s explosive impact across industries since its launch just over two years ago. The news industry is no exception. It now finds itself grappling with another technology-driven transformation – one that is already having a profound influence on how journalism is produced and consumed around the world. Undeniable opportunities have already emerged. AI can free up journalists’ time so they can focus more on their reporting and can also be used to make content more relevant for audiences, both of which are critical when the industry is fighting for sustainability and news avoidance is at record levels. At the same time, significant work remains to address ethical and cultural concerns and to ensure that trust – a key currency for journalism – is not undermined. Whilst the impact of AI has dominated recent conversation and research within the news industry, the discourse is often rooted in Western perspectives. Yet access to this technology differs worldwide, as do the problems faced by journalists and newsrooms. Recognising the need to address the imbalance, this report from the Thomson Reuters Foundation offers unique insight into how journalists across the Global South and emerging economies are using AI, and the distinct challenges they are up against. This report marks the first in our TRF Insights series, where the Foundation works with its global network of journalists – all alumni of our training programmes – to take timely pulse checks of major developments in the profession and provide data-led insights for the sector. Having trained and mentored thousands of journalists around the world for more than 40 years, our community is unique and diverse: our alumni are at different stages of their careers, they work for newsrooms of all sizes – some of which have been forced into exile – and report on a broad range of topics from financial corruption to barriers to healthcare to the climate emergency. The speed at which 4 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 we can analyse their responses to turn around a report – in this case, in under three months – helps to ensure that the findings and recommendations are as relevant as possible. It also means we can remain agile, as we can quickly apply learnings, or address gaps in support, through our programmatic work. Empowering journalists and newsrooms to navigate technological change is a key part of our work to bolster the resilience of independent media, which is critical to our mission to strengthen free, fair and informed societies. With AI here to stay, we have embedded a new module into our training programmes to support journalists in leveraging the benefits of this technology, developed a university curriculum for graduate journalists, and will soon be launching a newsroom adoption programme in South Africa, Kenya, Tanzania, Argentina and Uruguay. This research – made possible by the invaluable contributions of our alumni and its author Professor Damian Radcliffe, the Chambers Chair in Journalism at the University of Oregon – will further complement our work in this space. We also hope that it will be a useful resource for all those working to ensure that this technology is used in a responsible, ethical and equitable way across the news industry. We are only at the beginning of journalism’s AI journey, meaning it is still within our control as to how we shape the future. Now is the time to work together to ensure the right choices are made. And for that to happen, underrepresented voices must be heard. 5 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Executive summary REUTERS/Louafi Larbi 6 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 7 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Artificial Intelligence (AI) is already transforming the journalism industry and has the potential to dramatically change how news and information is consumed and created. Driven by new generative AI technologies, we may be witnessing the start of a new era of efficiency, creativity, and innovation in newsrooms. However, for all these potential benefits, the AI-age also presents significant challenges for journalists, particularly in areas such as ethics, equity, and access to digital technology. This new report from the Thomson Reuters Foundation (TRF) examines how journalists in the Global South and emerging economies1 are navigating this landscape by exploring their adoption of AI tools, as well as their hopes, concerns, and needs for a future where AI’s influence is likely to only be more prevalent and pervasive. Our research is grounded in results from a survey of more than 200 journalists in more than 70 countries around the world in Q4 2024. These respondents, alumni of TRF’s training programmes, shine a much-needed spotlight on the experience of journalism in the Global South and emerging economies, offering a timely snapshot of this rapidly evolving landscape. Main f ndings 01. AI adoption is widespread • Over 80% of survey respondents reported using AI tools in their journalistic work, with nearly half integrating these tools into their daily workflows, and nearly a further third using AI on a weekly basis. • Journalists are using AI tools for a diverse range of tasks, including drafting and editing content, research, transcription, translation, fact-checking, and idea generation, potentially helping journalists save time, reach new audiences, and work more efficiently. ChatGPT is the most widely used tool. 02. There are significant barriers to AI take-up • Despite a high level of adoption among our survey sample, only 13% of participants reported having an official AI policy in their workplace. Nearly eight out of ten (79.1%) respondents noted the absence of any clear company policy. • Similarly, nearly half (47.6%) of our respondents told us that their employers are neutral about AI integration. Around a quarter (22.6%) said their company actively 1 In this report, we have used the terms "Global South" and "emerging economies" to refer to developing countries and regions, primarily in Africa, Asia, Latin America, the Middle East, and parts of Eastern Europe. Many of these nations are seeing rapid economic growth, but may still face challenges in areas such as infrastructure and household incomes. We have used these terms to contrast with the more economically developed economies often referred to as the "Global North” or “Western Economies.” There is no standard definition agreed, although we note that the IMF uses the term “Emerging Market and Developing Economies,” and contrasts this with “Advanced Economies.” https://www.imf.org/en/Publications/WEO/Issues/2023/10/10/world-economic-outlook-october-2023 8 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 encourages it, with similar numbers (24.5%) being cautiously supportive. Only a small number (5.3%) of our sample noted that their company had banned or opposed the use of AI. 03. Users of AI in the Global South and emerging economies face unique challenges • Many leading AI models are primarily trained on English-language data from Western sources. This potentially risks creating an inherent bias in these systems, alongside exacerbating existing digital divides. • Nearly one in five respondents (18.3%) said they do not use AI. Among this cohort, lack of awareness and knowledge (55.3%) as well as insufficient training and support (52.6%) were the most cited barriers to usage. Lack of access (47.4%) and concerns about accuracy (45.0%) were also highlighted. Even among AI users, there appears to be a paucity of training, with the majority of those who use AI being self-taught (57.6%). 04. In some newsrooms, AI is already having a transformative impact • More than half of our sample (52.4%) said that AI had impacted their work in the past 18 months to either a “great extent” (12.4%) or a “large extent” (40%). Another third was more ambivalent, saying that it had “somewhat” of an impact (25.3%) or that it was too early to tell (9.4%). • That said, there are clear worries about long-term negative concerns, including AI’s impact on creativity and original reporting (54.3%), the erosion of critical thinking skills (51.4%) and the risk of increased misinformation (49%). 05. Journalists are divided about the future of AI technologies • A sizeable portion of survey participants (53.4%) expressed a high level of ethical worries about AI and its impact on the journalism industry, even though more than eight in ten are using AI technologies. Fewer than one in ten told us they were either “not concerned” (5.8%) or did not have a view on the matter at this time (2.4%). • As the adoption and use of this technology inevitably grows, survey participants expressed a desire for the creation of ethical guidelines, AI training, and regulatory frameworks designed specifically for the journalism industry. As a result of these findings, this report concludes with recommendations on page 47, offering actionable next steps for journalists, newsrooms, policymakers, platforms and funders. These proposals, based on survey insights and our own detailed analysis, highlight the tools, frameworks, and training needed if AI’s transformative potential is to be realised across the journalism industry in the Global South. 9 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Our research shows that journalists are cautiously optimistic about AI. However, they are wary about it being overused and worry about the human factor in reporting being diminished. Nevertheless, if adopted carefully and strategically, our survey suggests that journalists in the Global South and emerging economies believe AI can be a catalyst for strengthening and improving journalism, rather than it being a source of undue harm and disruption. REUTERS/Joe Skipper 10 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Introduction REUTERS/Yves Herman 11 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Artificial Intelligence (AI) is rapidly reshaping numerous industries, and journalism is no exception. The emergence of generative AI in the past couple of years has only accelerated this transformation, presenting challenges and opportunities, in terms of how the news is both produced and consumed. Against this backdrop, this new research from the Thomson Reuters Foundation (TRF) examines the adoption, and perception, of AI among journalists across the Global South and in emerging economies. This geographic focus matters because the take-up of AI by newsrooms, access to this technology, and attitudes towards it, varies around the world. As such, this report offers a fresh and valuable counterpoint to existing narratives which are often dominated by perspectives from Western economies and the Global North. Based on findings from a survey of alumni of TRF’s media training programs, this research features insights from a geographically diverse sample of 221 respondents across more than 70 countries. Conducted between 22 October and 3 November, 2024, our survey findings capture how journalists feel about the impact of generative AI on their industry. As our report demonstrates, AI tools are already transforming how news is researched, created, and distributed. It can do this by automating routine tasks, aiding translation and transcription, as well as enhancing data analysis, idea generation, and other storytelling efforts. At the same time, while AI offers opportunities to improve workflows and reach new audiences, the AI revolution also comes with inherent risks. This includes ethical challenges, and other issues related to the spread of misinformation, job displacement and technological inequalities. These concerns are particularly acute for journalists in the Global South and emerging economies, where unequal access to digital tools, training, and infrastructure, heightens the risk that certain newsrooms will be left behind. Compounding these challenges, many leading generative AI tools and models are primarily trained on data from Western and English-language sources. This can lead to an inherent bias within these systems, making them potentially less effective for communities and users in the Global South. Our survey respondents are alive to this complex landscape. For many of them, AI is a double- edged sword. On the one hand, it can streamline tasks, improve journalistic workflows, and help drive innovation and creativity. Yet, it is also seen as jeopardising core tenets of journalistic practice. Ethical implications, and the loss of a “human touch,” due to AI-assisted reporting, are among the issues that most troubled our survey sample. As news outlets experiment with, and adopt, AI, they must balance its benefits with efforts to mitigate the risks intrinsic to any new technology. In this report, we show how journalists around the world are walking this tightrope, by highlighting how they are embracing AI, navigating its complexities, and sharing their thoughts and fears about an AI-led future. We conclude with actionable recommendations, based on insights from journalists and editors worldwide. 12 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Snapshot of survey respondents The journalists and editors who participated in this survey bring a wealth of experience and diverse perspectives to the discussion of AI in journalism. In total, we received complete responses from 221 people working across the news industries in more than 70 countries across the Global South and emerging economies. The single largest group of participants identified themselves as reporters (43%), followed by editors (26.7%) and newsroom managers (8.6%). Just over four in ten respondents (41.5%) are aged 35 to 44, reflecting a group with significant industry experience. A further third are early-career professionals who fall into the 24 to 34 age bracket, with 25% of respondents aged 45 or older. In terms of gender representation, among respondents who disclosed this information, 51.2% identified as male, while 48.3% identified as female. We found that many reporters are actively embracing AI - in its many guises - in their work, with 81.7% of surveyed journalists using AI tools on a regular basis, despite limited organisational policies and formal training. It is their experiences with AI, as well as their hopes and concerns about this technology, that we feature throughout this new study. More details about the profile of our respondents can be found in the Appendix. There is a need for ongoing research offering global perspectives about the impact of AI and generative AI on the news media, including areas such as the Global South which risk being overlooked or underexplored. With that in mind, we hope that this report will contribute to that continuing conversation, offering timely insights that can be harnessed by newsrooms, journalism funders, media development organisations, and regulators, so that AI’s potential can be better understood and addressed, within the context of journalism in the Global South. 13 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Chapter 1: How is AI being used in newsrooms? REUTERS/Yannis Behrakis 14 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Artificial Intelligence (AI) has been used by newsrooms for more than a decade. Early efforts focused on a range of tasks including automating routine reporting, creating and tagging metadata, transcription, translation, and a number of other workstreams. The next iteration of this technology, generative AI, has emerged since late-2022 and quickly gone mainstream,2 beginning to transform multiple industries in the process. Journalism is no exception, with AI already starting to reshape how news is researched, produced, and consumed. This rapid technological evolution presents a mix of opportunities and challenges for journalists and newsrooms worldwide, which mirrors the differing perspectives that are seen throughout this report. On the one hand, tools like ChatGPT, Descript, and other AI-powered platforms are being integrated into journalistic workflows, promising a mixture of efficiency and creative augmentation. These technologies can help write articles, edit videos and podcasts, analyse data, create visualisations, and even generate content in multiple languages. This impact can be transformative. “It has made content generation faster and more efficient. AI has transformed how I conduct media research. With AI-powered tools, I can gather, analyse, and synthesize large volumes of data quickly. AI has also made editing more efficient, allowing me to ensure that content is error-free, concise and aligned with our brand’s tone.” Survey participant, Zimbabwe At the same time, AI technologies also raise profound questions about ethics, accuracy, and the role of human judgment in journalism. “ Whether using AI or not, journalist[s] must fact check, apply ethics and verify what AI has presented” Editor, Malawi Within this, concerns about misinformation, algorithmic bias and the erosion of critical journalistic skills are among the most pressing. “I am worried about plagiarism and misinformation . T here may be a blind belief in AI.” Reporter, Guatemala 2 See, for example: https://www.cnn.com/2019/08/24/politics/economy-us-workforce-automation/index.html, https://www.npr.org/ sections/alltechconsidered/2015/05/18/407648886/attention-white-collar-workers-the-robots-are-coming-for-your-jobs and https:// www.latimes.com/opinion/op-ed/la-oe-wright-robots-jobs-data-mining-20160328-story.html https://mediashift.org/2016/07/upsides-downsides-automated-robot-journalism/ https://gijn.org/stories/artificial-intelligence-demands-genuine-journalism/ https://techcrunch.com/2014/07/01/the-ap-is-using-robots-to-write-earnings-reports/ https://niemanreports.org/artificial-intelligence-newsrooms/ https://niemanreports.org/artificial-intelligence-newsrooms/ https://rjionline.org/news/five-different-ai-options-for-transcribing-audio/ https://alexandraborchardt.com/dont-mind-the-gap-automated-translation-could-revolutionize-journalism-but-how/ https://medium.com/damian-radcliffe/seven-ways-robots-are-being-used-by-publishers-and-newsrooms-22fc67b87761 https://www.vox.com/recode/2023/1/5/23539055/generative-ai-chatgpt-stable-diffusion-lensa-dall-e https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai https://www.ibm.com/think/insights/ai-in-journalism https://reutersinstitute.politics.ox.ac.uk/news/ai-and-journalism-whats-next https://nordmedianetwork.org/latest/news/journalism-and-ai-opportunities-and-threats/ https://edition.cnn.com/2019/08/24/politics/economy-us-workforce-automation/index.html, https://www.npr.org/sections/alltechconsidered/2015/05/18/407648886/attention-white-collar-workers-the-robots-are-coming-for-your-jobs https://www.npr.org/sections/alltechconsidered/2015/05/18/407648886/attention-white-collar-workers-the-robots-are-coming-for-your-jobs https://www.latimes.com/opinion/op-ed/la-oe-wright-robots-jobs-data-mining-20160328-story.html https://www.latimes.com/opinion/op-ed/la-oe-wright-robots-jobs-data-mining-20160328-story.html 15 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Alongside this, there are wider worries about job displacement, driven by long-standing media tropes that “robots are coming for your job,”3 and the impact of AI that is already being felt across major newsrooms4 and other industries. “I fear AI would render a huge majority of journalists jobless due to loss of work to AI and AI - powered systems.” Reporter, Kenya Although attitudes towards AI varied across our survey respondents, many expressed a cautious optimism about the potential AI may afford. This is not entirely surprising. After all, our respondents are self-selecting, having chosen to participate in a survey about AI, so their perspectives may reflect journalists who are more engaged with and aware of AI’s potential. “On the positive side, AI tools can streamline workflows, freeing journalists to focus on storytelling and in-depth reporting. Yet, there’s a real concern about over-reliance, especially if AI-generated content leads to uniformity or diminishes the human insight critical to good journalism.” Reporter, India Nevertheless, although this is not a representative sample, our survey responses also include the views of those less familiar with, or sceptical of AI, including non-users. Later in this report we dive into some of the challenges and fears that both groups shared with us. We begin, however, by exploring trends in AI usage, and how AI is being adopted in newsrooms across the Global South and emerging economies . 3 https://techcrunch.com/2014/07/01/the-ap-is-using-robots-to-write-earnings-reports/ 4 See: https://www.bbc.com/news/world-us-canada-52860247 and https://www.dw.com/en/ai-chatbots-replace-journalists-in-news- writing/a-65988172 for example. REUTERS/Joe Skipper https://www.bbc.co.uk/news/world-us-canada-52860247 https://www.bbc.co.uk/news/world-us-canada-52860247 https://www.businessinsider.com/chatgpt-jobs-at-risk-replacement-artificial-intelligence-ai-labor-trends-2023-02 https://techcrunch.com/2014/07/01/the-ap-is-using-robots-to-write-earnings-reports/ 16 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Journalistic functions and use AI adoption among the journalists that we surveyed is substantial, underscoring the importance of AI technologies in newsrooms around the world. More than eight in ten (81.7%) told us that they use AI in their work, with nearly half (49.4%) using these technologies daily, and a further 30.6% using it weekly. These journalists are leveraging AI for a variety of tasks, with improving content (55.3%), translation (51.8%), as well as research and inspiration (48.8%) being the top uses. Data analysis (42.9%) and fact-checking (37.6%) are also significant practices. REUTERS/Tim Wimborne Do you personally use AI in any form for your journalism? 17 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 “I can say for the past 18 months, AI has significantly enhanced my journalism. It has streamlined my research as it has confidently encouraged me to access complex data quickly and accurately, especially in areas like HIV research and environmental reporting.” Editor, Ghana The fact-checking potential of AI can support the needs of both journalists and consumers. One example of this is MyAIFactChecker. Described as “Africa’s first AI-powered fact- checking tool,” and available in seven different local languages, it was launched in 2024 by FactCheckAfrica, a Nigerian-based independent fact-checking platform. Users can verify news and information by inputting a query and then waiting for an answer to be returned. This includes details of the sources that have been searched. Image: Screenshot from the MyAIFactChecker website, 3 December 2024 Outside of these more popular uses of AI, tasks like transcription (34.1%), content creation (30.6% for text, 21.2% for multimedia), and design (18.2%) saw lower adoption rates. This may reflect the level of frequency with which individual journalists are undertaking these types of tasks, rather than being a reflection on AI technologies per se. https://www.myaifactchecker.org/ https://www.thecable.ng/factcheck-africa-launches-ai-initiative-journalism-fellowship/ https://factcheckafrica.net/ https://myaifactchecker.org/ 18 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 As AI tools develop, these numbers may also change, as technology enables journalists to take on tasks that were previously highly specialised. By way of an example, one journalist in Gambia, shared with us how they had used keywords in an AI program to design a logo. “I love the outcome,” they said. However, it’s clear that this type of work isn’t something that much of our sample does, or uses AI for, at this time. What functions do you use AI for in your work? 19 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 AI tools and their impact Not only are the journalists we sampled using AI on a regular basis, but they also acknowledge that it has often had made a major contribution to their work. “I use AI in designing still images for posters. I use AI in sound processing. I use AI in the translation of my subtitles. There is a proliferation of AI software in all categories: publishing, multimedia, visual infographics.” Reporter, Cameroon Among our survey respondents, more than half (52.4%) felt that AI’s impact had been discernible over the last 18 months. Just over one in ten (12.4%) said it had influence d their work “to a great extent,” in the past year while four in ten (40%) said it had impacted them “to a large extent.” “Thanks to AI I was able to quickly generate ideas in situations where it is impossible to brainstorm with colleagues and when it is difficult to start a task due to the “fear of a blank sheet”. In addition, it became possible to work with texts in different languages, analyse them and make extracts quickly and conveniently.” Manager, Armenia To what extent has your use of AI impacted your own work in the last 18 months? 20 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 One reason why some of our respondents have found AI to be so impactful can be seen in the breadth of ways that it can be used. As we have seen, journalists are harnessing AI for a range purposes. Examples cited by our sample include automatic language transcription, captioning and speech to text, as well as research and document analysis. “As a business reporter in Sub-Saharan Africa, AI has improved the depth at which I process data and do research for feature stories. This has led to richer context and depth in stories.” Reporter, Kenya Respondents also told us that AI had, in some cases, helped them save time too. According to one Editor in the Philippines, AI had “cut down workflow processes by 30 percent, particularly in analysis and basic proofreading.” Other participants highlighted similar time-saving benefits, using AI for idea generation, as well as a number of other activities including the drafting and editing of content. (It’s worth noting that outlets such as the Associated Press (AP) have been using automated reporting for more formulaic content, such as earnings reports, over the past decade.) “I use AI in text editing and idea generation, and this has saved me a lot of time. I find I develop initial drafts quicker, and AI has helped me edit and paraphrase.” Media m anager, Uganda ChatGPT is the most common tool cited by our sample, reflecting its leading status i n many markets around the world.5 As one manager in east Africa noted, “the paid version of ChatGPT has a bit of everything in it.” Another African-based respondent explained how “ChatGPT, developed by OpenAI, has become an essential companion in sub-editing my newspaper articles, offering a quick, reliable and efficient solution to refining my content before it reaches publication.” Though ChatGPT use is popular among our sample, journalists and managers are using a range of AI services, depending on the functions that they need. Our respondents are also using AI products like Grammarly, Canva, Gemini, Dall E, Otter, Meta AI, Midjourney and many others. One participant in India highlighted this plurality by sharing with us how “in my work, I’ve used ChatGPT for drafting emails and reports, Microsoft Power BI for data visualization and insights, and Grammarly for proofreading.” “These tools streamline communication, enhance data analysis, and improve document quality, allowing me to focus more on strategic tasks and ensure greater accuracy and efficiency,” they add. 5 See, for example, https://www.visualcapitalist.com/ranked-the-most-popular-generative-ai-tools-in-2024/ and https://www.statista. com/statistics/1548452/market-share-generative-ai-apps-downloads/ https://www.twipemobile.com/12-ways-journalists-use-ai-tools-in-the-newsroom/ https://www.inpublishing.co.uk/articles/how-were-using-ai-at-newsquest-24249 https://www.transformernews.ai/p/how-i-use-ai-as-a-journalist https://www.ap.org/the-definitive-source/announcements/automated-earnings-stories-multiply/ https://www.visualcapitalist.com/ranked-the-most-popular-generative-ai-tools-in-2024/ https://www.statista.com/statistics/1548452/market-share-generative-ai-apps-downloads/ https://www.statista.com/statistics/1548452/market-share-generative-ai-apps-downloads/ 21 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Non-users Despite the increasing deployment of AI in journalism, nearly one in five (18.3%) of respondents revealed that they do not use AI in their work. “I have not yet had the chance to confront the emergence of Artificial Intelligence in journalism.” Reporter, Benin The reasons for this included a lack of awareness and knowledge about the tools themselves, lack of training opportunities, limited access to AI tools, as well as issues of cost and connectivity, and additional concerns about the accuracy and reliability of these technologies. “I believe there is a need for more accessible AI tools tailored specifically for newsrooms in low-resource settings like ours.” Newsroom m anager, Sierra Leone If you do not use AI tools, what are the main reasons? 22 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 There is also a sense among some users that the benefits of this technology may have been over-hyped. As one e ditor based in Hong Kong put it : “the power and advances of AI often seem overblown to me.” “The quality of the content produced by AI seems to be lacking and ripe for manipulation,” they added, and “even simple spell checker tools like Grammarly do not have the same capabilities as a human editor.” The reticence that we see among some of our sample is understandable, not least because the journalism landscape is full of recent examples of technologies that were supposed to revolutionise journalism. From smartwatches to smart speakers and voice technology, through to blockchain, Web3, and immersive technologies like augmented and virtual reality - when it comes to how the news is created and delivered - these technologies have, so far, overwhelmingly underwhelmed. However, there is a sense that AI is different. Arguably, its use is already more prevalent and commonplace in newsrooms than several of these other new technologies. As a result, adoption may be at a tipping point, with the result that even its most ardent sceptics recognise this is a tech trend that may become impossible to ignore . “AI is the future, it is all around us, it is futile to fight it, it is all-powerful, it is inevitable. Or so I’m told. So, despite my great suspicion of AI - or should I say because of it - I have come to realise it is time to begin adequately familiarising myself with what I secretly fear as a formidable foe. Who knows, in it I might even find a favoured friend.” Editor, India REUTERS/Ali Khara https://www.journalism.co.uk/news/smartwatches-and-the-digital-future-of-news/s2/a555625/ https://reutersinstitute.politics.ox.ac.uk/news/new-report-future-voice-and-implications-news https://www.lse.ac.uk/media-and-communications/assets/documents/research/Blockchain-POLIS-Report-Jan-2019.pdf https://www.niemanlab.org/2021/12/journalism-goes-web3/ https://www.forbes.com/sites/solrogers/2020/02/06/is-immersive-technology-the-future-of-journalism/ 23 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Nevertheless, even AI’s advocates are alive to some of the challenges with this - still relatively - nascent technology. Misgivings around accuracy, for example, abound, and not just among non- users. As one e ditor from Bhutan reflected, “while I see the potential for AI to enhance reporting, I worry about the risk of misinformation and bias if these tools are not used responsibly.” “It’s essential to maintain our ethical standards and ensure that AI serves to support, rather than undermine, the trust we build with our audience,” they added. Meanwhile, a respondent from Ghana, while pointing to the benefits that AI can deliver for data- driven stories, also observed that “though generative AI tools are impressive, there are issues with bias [and the] production of misleading information.” “AI in journalism offers vast potential but requires careful handling. While AI can streamline research, detect trends, and personalize content, there’s a risk of compromising authenticity if overused. A primary concern is losing human judgment; machines may lack the nuanced understanding essential for ethical reporting.” Editor, Philippines These insights articulate the complex relationship that many journalists have with AI. Our survey suggests that a lot of journalists recognise its potential, however they also harbour doubts about possible negative effects on their profession. At the same time, the perceptions of non-users offer a helpful perspective into some of the ideological, technological and financial barriers that need to be overcome if AI adoption is to grow. These concerns need to be appreciated and addressed if efforts are to successfully bridge the gap between early adopters and their more hesitant colleagues. 24 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Chapter 2: Key challenges and risks REUTERS/Aly Song 25 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Although generative AI offers several creative and logistical possibilities for the journalism industry, the adoption of these technologies also brings with it a host of challenges. In this chapter, we explore the key concerns raised by our survey respondents, and the AI-related risks that they shared with us. These issues include ethical implications, questions of journalistic integrity, and navigating challenges related to digital divides. Lack of newsroom policies As we have seen, many of our survey respondents have rapidly embraced AI in their work, with the majority using AI, and nearly half using it daily. Yet, despite these levels of adoption, most journalists are utilising AI with little guidance. Only 13% of participants reported having an official AI policy in their workplace, with the vast majority (79.1%) noting the absence of any clear guidelines. This finding mirrors other research, which has found that news outlets are often operating without clear editorial procedures for using AI technologies. This is despite a growing body of published guidelines that media companies can adapt and learn from. Moreover, the presence of high profile “AI fails” from well-known western media outlets such as CNET, Sports Illustrated, MSN (featuring factual errors, tonally insensitive reporting, and fake bylines) has reinforced why newsrooms must have robust systems and editorial procedures in place. In many cases, these processes and principles are not just published internally, they are also shared with audiences,6 a move that is fundamental to building trust in the era of AI. 6 See, for example: https://www.thomsonfoundation.org/latest/paris-charter-on-ai-and-journalism-unveiled and https://www. niemanlab.org/2023/07/writing-guidelines-for-the-role-of-ai-in-your-newsroom-here-are-some-er-guidelines-for-that/ as well as https://www.nytimes.com/2024/10/07/reader-center/how-new-york-times-uses-ai-journalism.html Does your company have a published policy about the use of AI in its journalism? https://www.journalism.co.uk/news-commentary/wan-ifra-half-of-newsrooms-use-generative-ai-tools-only-a-fifth-have-guidelines-in-place/s6/a1038566/ https://generative-ai-newsroom.com/towards-guidelines-for-guidelines-on-the-use-of-generative-ai-in-newsrooms-55b0c2c1d960 https://generative-ai-newsroom.com/towards-guidelines-for-guidelines-on-the-use-of-generative-ai-in-newsrooms-55b0c2c1d960 https://journalistsresource.org/home/generative-ai-policies-newsrooms/ https://www.wired.com/story/cnet-published-ai-generated-stories-then-its-staff-pushed-back/ https://futurism.com/sports-illustrated-ai-generated-writers https://edition.cnn.com/2023/11/02/tech/microsoft-ai-news/index.html https://www.thomsonfoundation.org/latest/paris-charter-on-ai-and-journalism-unveiled/ https://www.poynter.org/commentary/2023/jouranlism-artificial-intelligence-ethical-uses/ https://www.thomsonfoundation.org/latest/paris-charter-on-ai-and-journalism-unveiled/ https://www.niemanlab.org/2023/07/writing-guidelines-for-the-role-of-ai-in-your-newsroom-here-are-some-er-guidelines-for-that/ https://www.niemanlab.org/2023/07/writing-guidelines-for-the-role-of-ai-in-your-newsroom-here-are-some-er-guidelines-for-that/ https://www.nytimes.com/2024/10/07/reader-center/how-new-york-times-uses-ai-journalism.html 26 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Within newsrooms, the lack of formal guidance leaves journalists to navigate AI adoption largely on their own. Subsequently, AI implementation is happening in a vacuum, with usage often driven by individual journalists rather than agreed organisational strategies and approaches. This internal policy vacuum inevitably leads to inconsistent practices, creating uncertainty and a lack of clarity for journalists and their editors. At the same time, it also increases the potential, whether accidental or intentional, for the misuse of AI technologies within a newsroom. “Some reporters have been misusing AI in going about their duties. The standard of AI- generated stories is also a concern.” CEO/Founder, Barbados For audiences, meanwhile, there is also a risk of confusion due to consuming content from different news sources, each of whom are potentially employing different approaches to AI in their work. This could significantly impact both actual and perceived levels of transparency, accuracy, and trust in journalism, as audiences grapple with inconsistent standards across different news sources. Absence of leadership and training The volume of newsrooms without AI policies is not altogether surprising when we explore the attitudes among newsroom leadership towards AI. According to our respondents, their management is typically rather ambivalent about these technologies. Nearly half (47.6%) of employers maintain a neutral stance on AI integration, while around a quarter (22.6%) actively encourage its use. A similar proportion (24.5%) are cautiously supportive. Only a small minority (5.3%) have banned or opposed AI technology, according to participants . Broadly speaking, what is your newsroom leadership’s attitude to AI? Has your employer provided or offered any formal training, resources or other support to use AI tools in journalism? 27 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 While there is openness to these technologies at many outlets, we also see a fair degree of caution, tentativeness and uncertainty. Some of this stems from lack of knowledge about how AI works and its possibilities. As one editor from Ethiopia acknowledged, “I need to know basic knowledge about AI and then we need to know how to use and exercise [it] in my corporation.” Similarly, a reporter from Kenya shared that “from my point of view, AI has not yet been fully or marginally embraced.” “I have found that this is due to a lack of knowledge, training and capacity of many media houses,” they added. Another reason for lack of leadership and clear policies on AI adoption may stem from a deliberate “wait and see” approach. It is easy to forget that generative AI tools are still relatively new, and some outlets may still be weighing the pros and cons of this technology, while others are forging ahead. “AI has the potential to completely reshape journalism, offering tools that can enhance efficiency and content creation. However, there’s uncertainty about whether this change will be entirely positive.” Editor, Pakistan Consequently, without a clear direction from senior leadership and in the absence of structured training opportunities, many journalists are taking matters into their own hands. A majority (57.6%) of AI users are self-taught, proactively engaging with technology, and leveraging online courses and guides (48.8%) to enhance their skills and knowledge. Other professional development opportunities, in the form of workshops and seminars, have been taken up by 40% of AI users. REUTERS/Thomas Mukoya 28 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 As we will discuss in the next chapter, there is an appetite among our sample for this to be addressed. As one manager from Lebanon told us, “I would love to get professional training on AI and journalism,” while another editor from Uganda hoped that “AI training [will be] incorporated in routine capacity-building sessions for journalists and even in journalism training in schools.” “I strongly feel that, with training, the fears about AI will be defeated by the benefits.” CEO and Founder, Zambia How do you currently learn to use AI? Check up to three most useful. Reduced transparency and accountability The emergence of AI raises fundamental questions about journalism’s core values, purpose and unique characteristics. Because of this, ethical concerns surrounding AI loom large, and the lack of guidance and training identified by our survey respondents may only be elevating this. “I hope AI-generated write-ups can be well distinguished from the ones written by human writers.” Reporter, Bangladesh Despite high levels of AI usage among our sample, more than half of our respondents (53.4%) nevertheless expressed high levels of concern about AI’s impact on ethics within the journalism industry. Notably, 20.2% reported being “extremely concerned,” while 33.2% were “very concerned” about AI’s ethical implications. Alongside this, additional worries expressed by our sample include the potential erosion of creativity and original reporting (54.3%), diminishing critical thinking skills (51.4%), and the risk of increased misinformation (49%). 29 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 What, if any, do you see as potential long-term risks to journalism regarding the increasing use of AI? (Select top three) As a reporter in Ethiopia put it: “My concern about AI in journalism is that it could stifle creativity and originality, potentially leading to job losses for journalists. As AI takes on more content creation tasks, there’s a risk that unique human perspectives and critical thinking may be overshadowed, diminishing the quality and authenticity of journalism.” Furthermore, a lack of clearly available guidance – published internally and externally – about the usage of this technology reduces transparency. That matters at a time when trust in journalism and the media is at, or near, record lows. Because of this, one reporter from Ghana advised, “as much as AI is here to make things ‘automated’ and easier, increased awareness about its use for journalism is needed to protect the integrity of the profession and practitioners.” With this in mind, it’s worth noting that research from the Reuters Institute shows that audiences overwhelmingly want journalists to disclose when they have used AI, especially for content creation. Failure to disclose AI use, therefore may potentially erode audience trust, especially if AI- generated content is not properly labelled or there are no clear standards for its use. https://www.statista.com/statistics/308468/importance-brand-journalist-creating-trust-news/ https://trustingnews.org/trusting-news-artificial-intelligence-ai-research-newsroom-cohort/ https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2024/public-attitudes-towards-use-ai-and-journalism https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2024/public-attitudes-towards-use-ai-and-journalism 30 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 “Clear ethical frameworks and regulatory guidelines for AI usage in newsrooms are essential to maintain transparency and audience trust, especially around AI-generated content.” Newsroom manager, United Arab Emirates Diminished creativity and the erosion of skills and jobs Original reporting, along with transparency, quality and accuracy, are among journalism’s core values. Although we have witnessed enthusiasm for the potential that AI might unlock, we also saw widespread apprehension about possible negative effects on some of the foundations and principles on which journalistic work is based. As we have seen, more than half of the surveyed journalists (54.3%) were concerned about AI’s long-term impact on creativity and original reporting. This was often expressed in terms of a deterioration in critical thinking. But it can be extended further to other core journalistic skills such as fact-checking, source verification, and in-depth analysis, as well as the risk of homogenised content. “Is AI about to kill what’s left of journalism?,” asked Rana Foroohar, a global business columnist and associate editor at the Financial Times, in a column published in May 2024. “I’m sceptical that artificial intelligence will benefit anyone aside from Big Tech in the short term,” she suggests. REUTERS/Jumana El Heloueh https://www.gulf-times.com/article/696083/opinion/experts-and-academics-over-reliance-on-ai-weakens-critical-thinking-skills https://www.ft.com/content/aa191322-13b1-4468-ab7b-431dfee2cc07 31 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Put simply, the conundrum that newsrooms face is rooted in finding the right balance between leveraging the benefits that AI may bring, without deskilling their journalists, and losing the key elements of people-powered reporting. “I fear the loss of originality and distinction. If everyone uses the same tools in the same way, we will face the same method, which may cause journalists to lose their distinction and uniqueness, as well as the personal human touch of the journalist.” Manager, Saudi Arabia Respondents also shared their fears that a reliance on AI would make content less empathetic and nuanced, especially when covering complex and sensitive subjects. After all, as one journalist reminded us, “journalism relies heavily on human empathy and understanding.” Developing these skills is fundamental to the journalistic craft. However, our sample indicated that younger journalists may miss opportunities to develop these foundational skills if AI technologies phase out entry-level roles. Similarly, the impact on job displacement, security of employment, and the wider risk of jobs being lost, are threats that journalists are very alive to in the age of AI. “There’s a fear that AI could replace human journalists, leading to job losses,” notes an editor in Tanzania. Although a common narrative is that artificial intelligence will make certain jobs redundant, one reporter in Bangladesh offered a different perspective. “If journalists get the job done with AI instead of talent, we may lose talented journalists,” they said. Others were equally sanguine. “I don’t fear AI taking my job,” one reporter wrote. “Every technology has its pros and cons, so journalists need to adapt.” REUTERS/Marco Bello https://en.arij.net/news/is-ai-taking-over-journalism-an-analysis-approach/ https://edition.cnn.com/2023/06/21/media/bild-germany-newspaper-ai-layoffs/index.html https://edition.cnn.com/2023/06/21/media/bild-germany-newspaper-ai-layoffs/index.html 32 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Perpetuating bias and misinformation For nearly half (49%) of our sample – the possibility of AI contributing to the spread of misinformation is also a major consideration. This may stem from the work produced by journalists themselves, given some of the flaws ( e.g. hallucinations) that are still inherent in many AI tools, as well as the ability of the public and bad actors to create false content. “Perhaps we can say that the level of searching for information has developed to some extent, but the profession of journalist remains the most prominent source of reliable information.” Editor, Tunisia In 2024, the World Economic Forum declared that “manipulated and falsified information is now the most severe short-term risk the world faces,” pointing to the role that generative AI was playing in supercharging this situation. These efforts may be intentional acts, deliberate pieces of disinformation created or amplified by AI, or AI - generated and manipulated content that people share without realising (or caring about) its accuracy. “I would like AI to learn to distinguish fakes from correct information and indicate the sources it used when composing the answer.” Respondent, Russia REUTERS/Ueslei Marcelino https://akademie.dw.com/en/generative-ai-is-the-ultimate-disinformation-amplifier/a-68593890 https://www.weforum.org/stories/2024/01/ai-disinformation-global-risks/ 33 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 AP has outlined examples of how AI technology has been used to manipulate discourse related to elections around the globe in 2024, while the Rest of the World website has produced a dedicated project, the 2024 AI Elections Tracker, which features examples of “the most noteworthy incidents of AI-generated election content globally.” Screenshot from Rest of World’s 2024 AI elections tracker, 11th December 2024 Beyond being used to create and spread misinformation, our survey respondents also shared concerns around the accuracy of AI technologies, as well as the biases inherent within them. For journalists in the Global South and emerging economies, this bias can manifest in several ways. For example, although our sample pointed to the value of these tools with transcription, AI technologies may struggle with local languages or dialects, potentially limiting their usefulness or accuracy. “AI algorithms can perpetuate biases present in the data they are trained on, leading to unfair or inaccurate reporting.” Editor, Tanzania That’s a view others share. In an in-depth interview published by Nieman Reports in early 2024, Casey Frechette, an associate professor and chair of the journalism department at the University of South Florida, points out: https://apnews.com/article/artificial-intelligence-elections-disinformation-chatgpt-bc283e7426402f0b4baa7df280a4c3fd https://restofworld.org/2024/elections-ai-tracker/ https://restofworld.org/2024/elections-ai-tracker/ https://niemanstoryboard.org/2024/02/16/journalism-ethics-artificial-intelligence-sources-fact-checking/ 34 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 “That’s the fallibility of the technology. It makes stuff up. It gets stuff wrong. It leaves out important caveats and context... In a similar vein, we all need to be concerned with the biases lurking in the datasets that AI models are trained on... reporters and editors need to be aware that many AI technologies have amplified, not diminished, the unfairness and misrepresentation that ail society and undermine the journalistic mission.” However, a few participants were enthusiastic about the potential of AI tools for countering bias, both in the present and future iterations of these technologies. “Artificial intelligence assists in reducing bias while interpreting data. Manually inspecting data and tracking down bias is quite difficult and tiring. AI programs and machine learning algorithms that are trained can help easily detect and put a break on biases in the world of journalism.” Respondent, Liberia That AI can be both the instigator, interlocutor and interpreter of bias, underlines the challenges that many journalists can face when using current AI technology. As a result, it is imperative that journalists understand the strengths and weakness of AI tools, and how to navigate these factors if they are to ensure that their reporting stays fair and accurate. In short, against this complex, fast-changing backdrop, it is essential, as one respondent from Indonesia remarked, that “journalists should not be afraid of AI, but instead must be able to master AI.” REUTERS/Dado Ruvic 35 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Chapter 3: Moving forward REUTERS/Thomas White 36 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 In this concluding chapter, we explore the hopes and concerns that journalists expressed about AI, and some of the core ideas they identified to support more widespread adoption in the near future. Within this, there is a cautious optimism, and sense of inevitability, about the growing impact AI will have on journalism. Nonetheless, there is also a feeling that if AI’s potential is to be fully and ethically realised, there remains a lot to be done. Attitudes towards AI and its impact on the future of journalism There is a deep paradox at the heart of how our survey participants feel about AI and its relationship with journalism. On the one hand, respondents recognise AI’s transformative potential and many of them are already using generative AI regularly for tasks like transcription, translation, and content generation. Yet, they also express caution about the take-up of these technologies and recognise that it is not without its shortcomings. This dichotomy is evident when we asked about how AI will impact the future of journalism. Just under half of our sample (42.3%) had a positive perspective, compared to one in five (20.2%) who held a negative view. A sizeable percentage (37.5%) indicated either a neutral stance (30.8%) or that they didn’t know (6.7%). Do you believe AI will have a positive or negative impact on the future of journalism? 37 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 These findings are especially interesting when we consider that 81.7% of journalists responding to our survey use AI in their work, with nearly half (49.4%) doing so daily, and a further third (30.6%) on a weekly basis. This tempered optimism suggests that many journalists see AI as a mixed blessing. “AI is only a threat if we do not learn how to use it. We should not allow it to take over critical thinking. While it can help with ideas, let us as journalists have [a] limit for what we feed in it.” Editor, Uganda While some of our respondents appear to be confident in their own abilities to harness AI, they may not be fully benefitting from its potential, and there is a risk that some of this confidence may be misplaced. Others, of course, lack any knowledge about AI altogether. Wherever journalists are on this spectrum, our survey indicates that the need for AI training and the clear editorial guidelines – findings that we outlined in Chapter 1 – is paramount. “My use of AI is limited due to my unfamiliarity with its proper application in my work. Additionally, I’m hesitant to overuse it without a clear understanding of its roles, best practices, and potential pitfalls.” Editor, Tanzania At the same time, our sample is also alert to the significant structural and strategic challenges that must be addressed for AI’s potential to be realised. These challenges, as outlined in Chapter 2, include tackling questions of accuracy, bias and ethical usage of AI tools. What this demonstrates is that you can be an early adopter of technologies like generative AI, while still being alive to its limitations. Many of our respondents are critically aware of the potential pitfalls associated with AI, at the same time as being active users. “AI can offer significant advantages – speed, data analysis and efficiency – that could enhance journalism’s reach and accessibility. However, it also presents challenges like misinformation and risks to journalistic integrity, making the overall impact a mixed but cautiously optimistic one.” Respondent, India A step change in training, AI capabilities and the establishment of agreed ethical standards may change attitudes towards AI and its role in journalism. It will be interesting to track sentiment over time, and to see if moves to elevate awareness, use and understanding of AI (efforts which do not gloss over discussion about limitations, ethical concerns, and the risks of over-reliance), will make people think differently about it. 38 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 The year ahead: a focus on regulation and ethical standards Reflecting on the next 12 months, survey respondents identified several key areas that require attention. Not surprisingly, many of these suggestions reiterated some of the ideas that we have already seen, such as the need for training and capacity building. In that vein, among the chief short-term challenges shown by our respondents were a lack of awareness and understanding of AI tools (50.6%), high costs (38.8%), and the speed with which this technology is changing (27.1%). However, the leading immediate issue, identified by more than half our sample (57.1%) focused on ethical concerns. “It would be good to see more discussion about ethical issues with the use of AI in the professional community. I am concerned that, in pursuit of audience attention and traffic, publications are neglecting journalistic standards, which consist primarily of providing truthful and honest information to the public.” Reporter, Russia REUTERS/Omar Sobhan 39 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 What are the main short-term challenges you face when using AI tools in your work? Increased training and the presence of newsroom guidelines can play a role in response to the ethical needs found by our survey. Without this, as one participant in Peru put it, “I’m afraid there will be… chaos and confusion.” Although newsrooms can determine their own guidelines, many of our respondents seem confident that regulatory frameworks will soon be forthcoming, and some of our sample would appear to welcome this. “I believe that it is necessary that there be regulation on the use of AI in journalism, to avoid unethical practices.” Editor, Mexico One reason participants may be open to more AI regulation stems from an acknowledgement that this is a complex landscape with lots of moving parts. Respondents expressed AI-related concerns across an assortment of topics including misinformation, data privacy, biases in AI platforms, and intellectual property considerations. “We certainly need laws to regulate the use of AI in both journalism and mainstream communication (influencers, blogs, for example).” Social media specialist, Brazil Addressing these questions is likely to require a multi-stakeholder approach, as many of these issues are outside the control of individual journalists and newsrooms, and it may be that market economics alone will not be enough to help resolve these concerns. https://www.brookings.edu/articles/can-journalism-survive-ai/ 40 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Subsequently, survey participants anticipate industry-wide policies and regulatory developments in the near future that are designed to tackle issues such as: • how journalistic content is used to train AI models • the creation of standards for disclosing when and where AI has been used (seen as fundamental to maintaining audience trust) • stronger protections for safeguarding sensitive information and sources when using AI tools. Alongside this, our respondents also hope that AI regulation may help with preserving, and possibly even enhancing, reputation and trust in journalism. By demonstrating clear ethical boundaries and transparent practices for AI use, news outlets may be able to differentiate themselves from other creators. By way of a potential solution, a manager from Africa recommended “some sort of accreditation for AI usage by media houses to distinguish us from other users.” “I would like to know how AI could be regulated in newsrooms to avoid plagiarism and lower the quality of ethics and content.” Reporter, Guatemala Interestingly, among non-users, and those who are wary about these technologies, it is possible that regulation may instil more confidence and willingness to get hands-on with AI. “For now, as there is no type of regulation, I avoid using AI as much as possible, mainly due to ethical issues. I think the tool can indeed help the journalist’s day-to-day work, however, the way it is, it leaves room for fake news.” Reporter, Brazil More widely, multiple journalists expressed a desire for regulatory interventions to assist with issues such as the cost and accessibility of AI, as well as development of tools designed to support fact-checking and investigative work. These are further barriers to take-up and adoption, especially in lower-income economies and smaller newsrooms. This is a long wish list, and regulation may not be a panacea. Progress to address issues via a regulatory route can be slow, and often both complex and complicated. Moreover, new standards and regulations are unlikely to be universal, which can further complicate adoption and roll out. “There will be new regulatory frameworks enacted, and they will vary around the world.” Respondent, Liberia http://dig.watch/resource/guidelines-on-the-responsible-implementation-of-artificial-intelligence-systems-in-journalism 41 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 We should also note that although respondents indicated a willingness for many concerns about AI to be remedied through regulatory means, they have little awareness of what regulations are in place in their own country or other markets. Fewer than a third (30.3%) of respondents indicated that they were aware of any legal regulations or guidelines governing the use of AI (e.g. around data protection, copyright infringement, etc) around the world. That drops to fewer than one in five (19.7%) saying they were aware of legal guidelines in their country of operation, perhaps reflecting the absence of these directives,7 as well as potentially poor, or confusing, communication of them. This demonstrates that if, and when, regulations are put in place, communication and enforcement may be significant obstacles to overcome. In these circumstances, clear information from policymakers, coupled with global and regional collaboration, may be essential for AI regulations to be understood and upheld. “At the rate the world is moving, in terms tech advancements, [it] could harm this profession. More regulations need to be put in place to protect our industry and the ethical conduct embedded in it.” Manager, Zambia 7 See, for example, : https://legalnodes.com/article/global-ai-regulations-tracker and https://www.whitecase.com/insight-our-thinking/ ai-watch-global-regulatory-tracker REUTERS/Adriane Ohanesian https://www.journalismliberty.org/tech-media-fair-compensation-frameworks https://www.washingtonpost.com/world/2023/09/03/ai-regulation-law-chia-israel-eu/ https://legalnodes.com/article/global-ai-regulations-tracker https://legalnodes.com/article/global-ai-regulations-tracker https://legalnodes.com/article/global-ai-regulations-tracker https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker 42 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Looking to the future: AI adoption goes increasingly mainstream Alongside a desire for greater clarity around ethical dimensions related to AI, as well as regulatory frameworks to manage usage, our survey respondents accept that the use of generative AI in journalism may only gather pace. Subsequently, many of the AI use cases that we have seen could become more prevalent in in the coming year. Image: Screenshot showing AI summaries of top stories on The Daily Maverick website, 14 December 2024 The Daily Maverick in South Africa, for example, has used generative AI to create summaries of their top stories, boosting readership in the process. They have also used AI technologies to improve their headlines, which has further helped grow their audience. In South America, the Peruvian news outlet Ojo Público has used AI to create new audio output, providing access to its investigative work in the Indigenous languages of Quechua, Aymara and Awajún, as well as Spanish. Meanwhile, virtual anchors can be seen “presenting” the news in countries such as India, and Venezuela, following in the footsteps of China, where AI hosts have been on screen since 2018. https://www.dailymaverick.co.za/summaries/ https://www.dailymaverick.co.za/ https://wan-ifra.org/2023/09/ai-use-cases-how-genai-summaries-are-boosting-daily-mavericks-readership/ https://wan-ifra.org/2023/09/ai-use-cases-how-genai-summaries-are-boosting-daily-mavericks-readership/ https://www.inma.org/blogs/ideas/post.cfm/daily-maverick-experiments-with-in-house-ai-solutions-while-learning-from-its-editors https://ojo-publico.com/ https://latamjournalismreview.org/articles/projects-use-generative-ai-to-bring-indigenous-languages-to-the-news-in-peru/ https://asia.nikkei.com/Business/Media-Entertainment/Rise-of-AI-newsbots-shakes-up-India-s-media-landscape https://edition.cnn.com/2024/09/18/americas/venezuela-retweets-ai-news-maduro-intl-latam/index.html https://www.theguardian.com/technology/article/2024/may/18/how-china-is-using-ai-news-anchors-to-deliver-its-propaganda 43 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 If they decide to embrace AI, news outlets can potentially benefit from a growing body of case studies like these,8 many of which stem from the Global South and emerging economies, thereby providing inspiration for newsrooms operating in similar contexts. In the process, it is hoped that AI will free up time for journalists9 to focus more on in-depth creative, original, and investigative work. But, this is not a given. “It is just as likely that any time savings will immediately be filled with new or additional demands,” cautions Felix Simon, a Research Fellow at the University of Oxford’s Reuters Institute. Nevertheless, as AI use in journalism potentially become more widespread, it only further reinforces the need for training, usage guidelines, greater direction from leadership, and the need to discuss fears about the negative influence of AI. Without proactively tackling these matters, the difficulties highlighted earlier in this report will only persist and escalate. Getting to grips with these challenges therefore must go hand-in-hand with a growing use of AI, so that the potential benefits of these technologies can be realised without exacerbating existing issues. 8 https://www.lse.ac.uk/media-and-communications/polis/JournalismAI/Case-studies - which links to https://airtable.com/ appdaeDwFizD4RK0u/shrKhe7Js48HvBhmG/tblBcSZESOAuy5Q9A 9 See, for example: https://www.techinasia.com/about/use-of-ai REUTERS/Victor Fraile https://www.lse.ac.uk/media-and-communications/polis/JournalismAI/Case-studies https://www.lse.ac.uk/media-and-communications/polis/JournalismAI/Case-studies https://www.techinasia.com/about/use-of-ai https://journalism.columbia.edu/news/tow-report-artificial-intelligence-news-and-how-ai-reshapes-journalism-and-public-arena https://www.lse.ac.uk/media-and-communications/polis/JournalismAI/Case-studies https://www.lse.ac.uk/media-and-communications/polis/JournalismAI/Case-studies https://airtable.com/appdaeDwFizD4RK0u/shrKhe7Js48HvBhmG/tblBcSZESOAuy5Q9A https://airtable.com/appdaeDwFizD4RK0u/shrKhe7Js48HvBhmG/tblBcSZESOAuy5Q9A https://www.techinasia.com/about/use-of-ai 44 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Conclusion and looking ahead REUTERS/Molly Darlington 45 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 The adoption of AI in journalism is one of the most significant technological shifts in the industry’s recent history, and this revolution has only just begun. An AI-led transformation is already reshaping how journalists research, create, and deliver content, unlocking efficiencies and sparking opportunities for creativity and innovation. In turn, AI is already beginning to redefine how, and where, some audiences look for news and information. We know that further change is coming, even if the extent of it is hard to estimate right now. As such, the future of journalism in the AI-age is rich with potential but fraught with complexity. Moreover, AI take-up by the wider population is also likely to affect journalism’s financial health. Even though it is difficult to predict the magnitude of this, AdWeek estimates that AI-driven search could shrink search traffic to news outlets by 20%-60%, reducing advertising revenues by “as much as $2 billion annually across the publishing industry.” Although there is mutual dependency between AI companies and newsrooms – with journalists increasingly using AI technology, and platforms relying on journalistic output to train their AI models – the benefits of AI for journalism remain uneven and uncertain. That is especially true for news outlets in the Global South, who can find it hard to get a seat at the table, and who often find themselves waiting in line behind their western peers, when it comes to discussions and partnerships10 with technology providers. To offset this, and to ensure that newsrooms and journalists in the Global South are not left behind, stakeholders such as news outlets, journalism funders, media development organisations and regulators, will need to prioritise activities grounded in five key areas: 01. Training and skills development to enable journalists to harness AI’s potential. 02. Ethical frameworks which make AI tools more inclusive and less biased. 03. Active partnerships between newsrooms, developers, platforms and funders to ensure that AI tools can best serve journalism and wider information needs. 04. Promoting transparency throughout the entire process – from developments by tech companies, to implementation in newsrooms and communication to audiences – so that openness is at the heart of how AI tools are used and created. 05. Addressing digital divides to deliver equitable access to AI technologies across diverse newsrooms and regions around the world. 10 See, for example: https://petebrown.quarto.pub/pnp-ai-partnerships/, https://digiday.com/media/2024-in-review-a-timeline-of-the- major-deals-between-publishers-and-ai-companies/ and https://pressgazette.co.uk/platforms/news-publisher-ai-deals-lawsuits- openai-google/ https://pressgazette.co.uk/comment-analysis/ai-revolution-for-news-publishers-is-only-getting-started/ https://www.adweek.com/programmatic/googles-gen-ai-search-threatens-publishers-with-2b-annual-ad-revenue-loss/ https://niemanreports.org/the-battle-over-using-journalism-to-build-ai-models-is-just-starting/ https://petebrown.quarto.pub/pnp-ai-partnerships/ https://digiday.com/media/2024-in-review-a-timeline-of-the-major-deals-between-publishers-and-ai-companies/ https://digiday.com/media/2024-in-review-a-timeline-of-the-major-deals-between-publishers-and-ai-companies/ https://petebrown.quarto.pub/pnp-ai-partnerships/, https://digiday.com/media/2024-in-review-a-timeline-of-the-major-deals-between-publishers-and-ai-companies/ and https://pressgazette.co.uk/platforms/news-publisher-ai-deals-lawsuits-openai-google/ https://pressgazette.co.uk/platforms/news-publisher-ai-deals-lawsuits-openai-google/ https://pressgazette.co.uk/platforms/news-publisher-ai-deals-lawsuits-openai-google/ 46 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 By focusing on these activities, which we expand on in more detail in the recommendations section of this report, AI can help transform and elevate journalism, strengthening, rather than undermining it. That won’t be easy, but if managed carefully and ethically, AI may contribute to efforts designed to make journalism more innovative, inclusive, and efficient. Delivering on this promise will not be easy. These challenges are significant, but so are the opportunities. This duality is understood by our respondents, many of whom are cautiously enthusiastic about the future of journalism in an AI world. We hope to be able to report in the future that their optimism was not unfounded. The AI-age has just begun. TRF and our alumni, look forward to helping shape it. “I think it’s the most valuable innovation of our time. We just have to cautiously use it.” Reporter, Zambia REUTERS/Kacper Pempel 47 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Recommendations for stakeholders REUTERS/Aly Song 48 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 The successful integration of AI into journalism in the Global South requires a multi-stakeholder approach. No single organisation, platform, funder or regulator can go it alone. Subsequently, based on our detailed analysis of the findings from the 2024 TRF Alumni AI Survey, we conclude this report with recommendations for five key groups: journalists, newsrooms, policymakers, journalism funders and media development organisations, as well as platforms and developers. Our recommendations are designed to foster the responsible and ethical development of AI and its integration into journalistic work in the Global South and emerging economies, although they are not exclusive to those regions. Recommendations for journalists What’s needed? Ethical knowledge and practical skills, understanding of AI opportunities and limitations. Key recommendations: 01. Invest in your own AI mastery: Proactively seek AI training and skill development opportunities, including through self-learning, to better understand the strengths and weaknesses of AI technologies. 02. Experiment with AI tools: to enhance efficiency and creativity, without compromising critical thinking, creativity, and fact-checking skills, or becoming over-reliant on these tools, especially for complex and sensitive topics. 03. Understand the ethical implications of AI: This includes advocating for policies and practices that protect journalistic integrity, while also ensuring fairness and accuracy, and promoting trust with audiences. 04. Share experiences and best practices: Knowledge sharing can engender collective learning, with a focus on using AI effectively and responsibly, so that it enhances journalistic output and workflows. “ I am a strong believer in originality and critical thinking and if using AI can ease the work of journalists without getting rid of originality or increasing misinformation, then it should be welcome in the newsroom.” Respondent, Gambia 49 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Recommendations for editors and newsroom leaders What’s needed? Clear AI policies, ethical frameworks, investment in AI tools and training. Key recommendations: 01. Establish clear AI policies: Newsrooms should develop frameworks outlining acceptable and responsible uses of AI. This will help ensure consistent AI practices, as well as promote transparency and accountability internally and with audiences. 02. Provide training for all: Journalists and editors need access to resources and guidance for integrating AI into their work. Training should cover AI’s strengths and weaknesses, use cases, ethical considerations and implementation guidelines. 03. Foster a culture of AI literacy: Leaders can actively nurture AI-focused learning and experimentation in their newsrooms, and with partners such as other news outlets and funders. Collaborate to create shared, consistent, ethical principles and best practices, which may help to engender or maintain trust in journalism. 04. Bring your audience with you: Educate the public about AI, through transparency about AI’s role in your journalism and how information is created and shared. Highlight skills and knowledge needed to navigate an increasingly AI-driven world. “Newsrooms should start talking about AI and explain their policies about AI to the journalists. In most newsrooms, what I see is, AI is considered a taboo word. But, newsrooms should take the initiative to permit the use of AI within an ethical framework.” Editor, Sri Lanka REUTERS/Marco Bello 50 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Recommendations for media development organisations and funders What’s needed? Investment in training, and promoting ethical practices, inclusivity and innovation. Key recommendations: 01. Invest in AI training and capacity-building tailored to the Global South: Lack of training and understanding is a key barrier to using AI. To build capacity in practical applications of AI and ethical knowledge, more support is required, especially in smaller markets, newsrooms and among speakers of less widely spoken languages. 02. Ensure that freelancers are not left behind. Invest in programmes that improve the AI skills, knowledge and access for journalists who are self-employed. 03. Support initiatives which develop and share AI tools to counter digital divides: Concerns about the cost of AI technologies are very real. Addressing this could involve support for development of open-source AI tools, or subsidies for software that may be unaffordable for some newsrooms in the Global South. 04. Fund pilot projects and research specific to journalism in the Global South: This might involve exploring how AI can enhance investigative reporting, fact-checking and combating misinformation and efforts to reach audiences via new formats. 05. Facilitate knowledge sharing and networking: Design platforms, communities, and other forums, so that journalists in the Global South can more easily share best practices, case studies and resources, with one another. “We hope that the tools available for training will be available, especially in developing countries and undeveloped countries such as African countries.” Editor, Tunisia REUTERS/Navesh Chitrakar 51 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Recommendations for policy makers What’s needed? Tackling digital divides, creation of regulatory frameworks, investment in local AI solutions. Key recommendations: 01. Develop regulatory frameworks that support and protect journalism: Balance innovation with ethical considerations and public trust, targeting issues related to topics such as data privacy and protection, transparency and ethical usage, safeguarding against the spread of misinformation, preserving copyright and intellectual property and tackling potential biases in AI systems. 02. Bridge resource gaps and promote capacity building: Addressing disparities in AI access and skills may require intervention in under-resourced regions. This might include innovation grants and pilot projects, investments in infrastructure and AI tools tailored to less widely spoken languages, as well as training programmes. 03. Encourage cross-sector collaboration: Newsrooms, NGOs, platforms and academic institutions can work more closely to support AI-related research, development and learning. Issues that would benefit from being addressed, and better understood in a Global South context include bias mitigation, data privacy, and AI’s impact on newsgathering and production. “I anticipate regulatory developments to ensure transparency in AI-driven journalism and the promotion of unbiased content. My belief is that ethical AI usage, combined with these technological advancements, could strengthen journalistic integrity and accountability.” Editor, Ghana REUTERS/Adriane Ohanesian 52 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Recommendations for platforms and developers What’s needed? Journalism-focused tools, platform-publisher partnerships, confronting technology gaps. Key recommendations: 01. Create more AI tools designed for journalistic needs and workflows: Adopt a more journalism-centric approach in lower-income economies by developing products to help uphold core journalistic principles like accuracy, fairness and verification, and not just content generation or automation technologies. 02. Prioritise transparency: To build trust and accountability with journalists and the public, ensure AI systems disclose their datapoints and provide clear explanations for AI outputs, particularly in areas like fact-checking and source verification. 03. Address biases in AI models: Invest in local solutions, especially those affecting underrepresented languages and communities, working alongside journalists to support them, rather than replacing core journalistic functions and content. “AI biases – often unnoticed – can skew narratives or reinforce stereotypes, affecting public perception. However, I’m optimistic that, with balanced integration, AI can support rather than replace journalists, enabling deeper, data-driven reporting. The key will be setting clear boundaries, fostering transparency, and ensuring that AI remains a tool rather than a storyteller.” Editor, Philippines REUTERS/Kenny Wu 53 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Appendix 1: Methodology and participant sample This report serves as a snapshot of the state of generative AI in newsrooms across the Global South and emerging economies, at the tail end of 2024. The study is based on findings from an online survey conducted between 22 October and 3 November, 2024. Respondents are alumni of the journalism training programmes run by the Thomson Reuters Foundation (TRF), and they were recruited via an email sent to people who had participated in a TRF training course who had opted to receive ongoing communication from TRF. The online survey was available in English, Arabic, French, Russian, Spanish and Portuguese. We received 221 complete responses to our online survey, with respondents based in more than 76 countries and territories around the world. Survey job titles More than four-in-ten (43%) of respondents identified themselves as reporters, with a further quarter (26.7%) stating that they were editors. The remaining 30% of survey participants are employed in a range of different roles from newsroom operations (8.6%) through to social media (3.2%). A small percentage (4.5%) are not primarily employed in a media or journalism role. Journalism experience Overall, our survey respondents are experienced newsroom practitioners. Just over two-thirds (66.8%) have been working in journalism for more than a decade. Nearly a further quarter (23.6%) have five to nine years’ experience. Fewer than 10% of respondents have been working in journalism for four years or less. As a result, our survey respondents are not only experienced journalism professionals, but they are also likely to have already experienced multiple waves of digital disruption. In the past two decades, the industry has been impacted by developments such as the advent of social media, mobile phones and the move to many outlets becoming “digital first.” These experiences inevitably shape the perspective that they bring to this survey; after all, many of these professionals have weathered previously seismic shifts in 54 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 the industry’s structure, the technology used to create and consume journalism, as well as changes in audience expectations and habits. This gives them a unique vantage point to assess the arrival of generative AI. Gender and age The bulk of participants in this study are likely to be in the middle of their media careers. More than four in ten (41.5%) are aged 35 - 44 years old, while a third are aged 24-34. A further quarter of our sample are aged over 45, which nearly one in five (19.3%) aged 45-54. The gender of our survey respondents was almost equally spit, with 51.2% of respondents sharing this demographic data with us identifying as male, and 48.3% identifying as female. Areas of media focus Our survey respondents come from a wide range of different media outlets. We asked them to identify the primary output of their company, with 63% saying that this was online text, just over half (50.5%) having a print focus and the same number pointing to online multimedia. Broadcast TV was represented by nearly a quarter (23.6%) of our sample, with radio being a main channel for 16.3% of our survey participants. Is your media primarily focused on: 55 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Survey country mix The 76 countries and territories represented in the survey included India, Kenya, Nigeria, Ghana, Uganda, Brazil, Russia, Bangladesh, Cameroon and Ethiopia. Collectively, respondents from 10 countries made up 86 of the 203 participants who shared where they are based. The remaining 117 respondents who shared this information came from places as diverse as Tanzania and Zimbabwe, Mexico and Guatemala, through to the Philippines and Nepal, Tunisia and Lithuania. What country are you based in? 56 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Appendix 2: Survey questions 01. What is your current job function (check the item that best matches your function)? • Reporter • Editor • Researcher • Newsroom/media manager/operations • Audience engagement • CEO/Founder • Social media • Not in a journalism or media profession • Other - please specify 02. Please rank the following priorities from your personal perspective when carrying out your job, with 1 being the highest priority and 7 being the lowest priority. • Accuracy in reporting • Ethical journalism • Investigative depth • Source protection • Meeting deadlines • Audience engagement • Understanding and using AI 03. Broadly speaking, what is your newsroom leadership’s attitude to AI? Has your employer provided or offered any formal training, resources of other support to use AI tools in journalism? • Actively encouraging/providing training • Cautiously supportive • Neutral • Opposes • Banned 57 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 04. Does your company have a published policy about the use of AI in its journalism? • Yes • No • Don’t know 05. Do you personally use AI in any form for your journalism? This could include AI for content creation, research, translation, SEO etc • Yes • No 06. If you do not use AI tools, what are the main reasons? (Check all that apply) • Lack of awareness or knowledge about these tools • Insufficient training or support • Lack of access to necessary technology • Concerns about accuracy and reliability • Ethical concerns • Prefer traditional methods • Cost or budget constraints • Fear of job displacement • Other (please specify) 07. How frequently do you personally use AI tools in your journalism work? • Daily • Weekly • Monthly • Less than once per month 58 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 08. What functions do you use AI for in your work? (Select all that apply) • Improving content (e.g. editing) • Translation • Inspiration / Idea generation • Research • Data analysis • Fact-checking • Transcription • Content creation - text • Content creation – multimedia • Design • Audience engagement / Audience analysis • Other - please specify 09. To what extent has your use of AI impacted your own work in the last 18 months? • To a great extent • To a large extent • Somewhat • Little • Not at all • Can’t tell yet / too early to say 10. Can you share some examples? (Free text) 11. What AI tools have you used in your work? (Free text) 12. How do you currently learn to use AI? Check up to three most useful. • Playing with tools • Online courses / guides • Workshops / webinars / seminars • From my company / colleagues • Community forums • I haven’t learnt about AI yet • Other - please specify 59 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 13. What are the main short-term challenges you face when using AI tools in your work? (Select top three) • Ethical concerns • Lack of awareness or understanding about the tools • High cost • Rapid evolution of the technology • Resistance or lack of interest from colleagues / managers • Unreliable technology / access to internet • No time to learn • Other - please specify • No challenges 14. What, if any, do you see as potential long-term risks to journalism regarding the increasing use of AI? (Select top three) • Loss of creativity and original reporting • Erosion of critical thinking skills • Increased misinformation • Loss of journalistic integrity • Job security • Widening gap between AI-equipped and traditional newsrooms / journalists • Legal liabilities • Other - please specify • No risk 15. Do you believe AI will have a positive or negative impact on the future of journalism? • Very positive • Somewhat positive • Neutral • Somewhat negative • Very negative • Don’t know 60 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 16. Please give examples of how? (free text) 17. Are you concerned about the ethical implications of using AI across the journalism industry? • Extremely concerned • Very concerned • Moderately concerned • Slightly concerned • Not concerned • Don’t know 18. Are you aware of any legal regulations or guidelines that govern the use of AI (e.g. around data protection, copyright infringement etc): Internationally? • Yes • No • Don’t know In your country? • Yes • No • Don’t know 19. What do you think has been the most significant development in AI related to journalism over the past year? (Free text) 20. What changes or developments do you hope to see in the coming year, regarding AI and journalism? This could be related to new tools, ethical frameworks, training, or AI integration in your newsroom, potential technological advancements, regulatory changes, or shifts in AI usage. (Free text) 21. Would you like to share any other observations or comments on your use, hopes or fears of AI in journalism? (Free text) 61 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 22. What’s your gender? • Male • Female • Prefer not to say 23. What’s your age? • 18 to 24 • 25 to 34 • 35 to 44 • 45 to 54 • 55 to 64 • 65 to 74 24. What is your nationality? 25. What country are you based in? 26. How long have you been working in the journalism profession? • Under a year • 1 to 4 years • 5 to 9 years • 10+ years 27. Is your media primarily focused on: (Select up to three, if you work equally across formats. If a freelancer, select the option most appropriate to your regular work.) • Online text • Print • Online multimedia • Social • Broadcast TV • Broadcast Radio • Other - please specify 62 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Appendix 3: Suggested reading & resources Below are 15 recommended reports and articles for those interested in exploring this topic further, many of which are cited in this report. While not exhaustive, this list serves as a helpful starting point for delving into many of the key ideas discussed in this report. Given the rapidly evolving landscape, we have included publication dates which may help guide readers on where to begin. Mapping the Finances and Ownership of AI Companies Used by Journalists, Sydney Martin. Edited by Marius Dragomir. Published by Media and Journalism Research Center. (November 2024) AI and the Future of Journalism: An Issue Brief for Stakeholders, Anya Schiffrin. Published by UNESCO. (October 2024) AI in Action: 10 Global Use Cases. Published by WAN-IFRA (World Association of News Publishers). (May 2024) AI Survey Report on the Adoption of Artificial Intelligence (AI) by Journalists in Zambia. Internews. (April 2024) Generative AI in Journalism: The Evolution of Newswork and Ethics in a Generative Information Ecosystem, Nicholas Diakopoulos, Hannes Cools, Charlotte Li, Natali Helberger, Ernest Kung, Aimee Rinehart, edited by Lisa Gibbs. Published by AP. (April 2024) Artificial Intelligence in the News: How AI Retools, Rationalizes, and Reshapes Journalism and the Public Arena, Felix M. Simon. Published by the Tow Center for Digital Journalism. (February 2024) Journalism, Media, and Technology Trends and Predictions 2024, Nic Newman. Published by Reuters Institute for the Study of Journalism at Oxford University. (January 2024) Seven steps for a successful AI strategy, Damian Radcliffe. Published by Digital Content Next. (September 2023) Generating Change: A global survey of what news organisations are doing with AI, Charlie Beckett and Mira Yaseen. Published by The London School of Economics and Political Science. (September 2023) https://journalismresearch.org/mapping-the-finances-and-ownership-of-ai-companies-used-by-journalists/ https://unesdoc.unesco.org/ark:/48223/pf0000391214 https://wan-ifra.org/2024/05/wan-ifra-publishes-5th-report-on-ai/ https://internews.org/resource/ai-survey-report-on-the-adoption-of-artificial-intelligence-ai-by-journalists-in-zambia/ https://internews.org/resource/ai-survey-report-on-the-adoption-of-artificial-intelligence-ai-by-journalists-in-zambia/ https://www.researchgate.net/publication/379668724_Generative_AI_in_Journalism_The_Evolution_of_Newswork_and_Ethics_in_a_Generative_Information_Ecosystem?channel=doi&linkId=6614c3cc43f8df018de7606c&showFulltext=true https://www.researchgate.net/publication/379668724_Generative_AI_in_Journalism_The_Evolution_of_Newswork_and_Ethics_in_a_Generative_Information_Ecosystem?channel=doi&linkId=6614c3cc43f8df018de7606c&showFulltext=true https://www.cjr.org/tow_center_reports/artificial-intelligence-in-the-news.php https://www.cjr.org/tow_center_reports/artificial-intelligence-in-the-news.php https://reutersinstitute.politics.ox.ac.uk/journalism-media-and-technology-trends-and-predictions-2024#header--7 https://digitalcontentnext.org/blog/2023/09/28/seven-steps-for-a-successful-ai-strategy/ https://static1.squarespace.com/static/64d60527c01ae7106f2646e9/t/656e400a1c23e22da0681e46/1701724190867/Generating+Change+_+The+Journalism+AI+report+_+English.pdf 63 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 ChatGPT and the Global South: how are journalists in sub-Saharan Africa engaging with generative AI?, Gregory Gondwe. Published in the journal Online media and Global Communication. (June 2023) AI, Journalism, and Public Interest Media in Africa, George Ogola. Published by International Media Support (IMS). (June 2023) Journalists Toolbox. Features AI resources across more than 30 categories. AI Unlocked. A seven-part webinar series, from WAN-IFRA (World Association of News Publishers) and FATHM, a London-based news consultancy, and creative studio. The Fix AI Course. Eight week email course offering advice on AI for news leaders. JournalismAI Discovery. A six module course, launching in February 2025 in Spanish, by the London School of Economics and Political Science (POLIS), with support from the Google News Initiative (GNI). https://www.degruyter.com/document/doi/10.1515/omgc-2023-0023/html https://www.degruyter.com/document/doi/10.1515/omgc-2023-0023/html https://www.mediasupport.org/publication/ai-journalism-and-public-interest-media-in-africa/ https://journaliststoolbox.ai/ https://wan-ifra.org/ai-unlocked/ https://newsletter.thefix.media/ai https://www.journalismai.info/programmes/discovery 64 THOMSON REUTERS FOUNDATION INSIGHTS REPORT 2025 Appendix 4: About the author Damian Radcliffe is a journalist, educator, researcher, and professor based at the University of Oregon. He holds the Chambers Chair in Journalism and is a Professor of Practice, an affiliate faculty member of the Department for Middle East and North Africa Studies (MENA) and the Agora Journalism Center, and a Research Associate of the Center for Science Communication Research (SCR). He is an expert on digital trends, social media, technology, the business of media, the evolution of present-day journalistic practice. Damian is a three-time Knight News Innovation Fellow at the Tow Center for Digital Journalism at Columbia University, an Honorary Research Fellow at Cardiff University’s School of Journalism, Media and Culture Studies (JOMEC), and a Life Fellow of the Royal Society for the Encouragement of Arts, Manufactures and Commerce (RSA). In Spring and Summer 2023 he was a Visiting Fellow at the Reuters Institute for the Study of Journalism, a Visiting Academic Associate at Green Templeton College, at the University of Oxford. With nearly 30 years of experience in the media industry, Damian has worked in editorial, strategic, research, policy and teaching roles in the USA, Middle East and UK. This includes roles in all media sectors (commercial, public, government, regulatory, academic, and nonprofit/civil society) and all platforms (print, digital, TV and radio). https://www.linkedin.com/in/damianradcliffe/ https://journalism.uoregon.edu/directory/faculty-and-staff/all/damianr https://towcenter.columbia.edu/content/damian-radcliffe%E2%80%9D%20target= https://profiles.cardiff.ac.uk/honorary/damian-radcliffe https://damianradcliffe.wordpress.com/2008/03/14/rsa/ https://reutersinstitute.politics.ox.ac.uk/people/damian-radcliffe https://reutersi