Navigating the Algorithmic Frontier: How News Organizations are Crafting, Implementing, and Evolving AI Policies
As the world grapples with the accelerating ascent of artificial intelligence and fundamental questions about what guardrails should exist, news organizations find themselves at a critical juncture, tasked with defining their own operational parameters for this transformative technology. While past waves of innovation, such as blogging and social media, prompted newsrooms to draft new policies primarily concerning content distribution and engagement, the current AI revolution presents a distinctly different and more profound challenge. This technology is not merely altering how journalistic work is disseminated but fundamentally reshaping how it can be produced, from initial data gathering to final content generation. The stakes are exceptionally high: a newsroom that permits AI-generated inaccuracies or "hallucinations" to slip into production risks irreparable damage to its most valuable asset—its reputation and public trust.
The urgency for clear, actionable AI policies is underscored by a consensus among media leaders that this technological shift is unprecedented. Unlike previous innovations that primarily impacted workflow or audience interaction, AI delves into the very core of journalistic creation. The ability of generative AI to produce text, images, and even video raises complex ethical questions about authorship, authenticity, and accountability. This profound shift necessitates a strategic and iterative approach to policy development, moving beyond simple guidelines to comprehensive frameworks that integrate ethics, training, and continuous oversight.
To better understand how newsrooms can effectively write and implement AI policies, the Columbia Journalism Review (CJR) engaged with Anika Collier Navaroli, a distinguished expert in policymaking and a veteran of the Trust and Safety teams at Twitter and Twitch. Navaroli, who now directs the Craig Newmark Center for Journalism Ethics and Security at Columbia Journalism School, emphasizes that AI policy development often receives insufficient attention until a crisis strikes. "Nobody celebrates a policy win, because that’s just a regular good day," she noted, highlighting the proactive nature required to prevent potential pitfalls rather than reacting to them.
Further insights were gathered from a diverse array of experts across media organizations and journalism support groups, revealing both common strategies and unique challenges. Jane Barrett, the head of product at Reuters, stressed the need for agility: "The AI world moves incredibly fast, so all of us publishers need to keep reviewing our policies." This sentiment was echoed by Sashka Koloff, the standards editor for the Australian Broadcasting Corporation (ABC), who underlined the importance of policies aligning with "audience expectations, making sure that they actually work in practice." Tess Jeffers, the head of newsroom AI and data at the Wall Street Journal, articulated a forward-looking mindset, stating, "It was very important for us to lead with excitement rather than fear," suggesting a proactive embrace of AI’s potential while acknowledging its risks.
A universal theme emerged from these discussions: merely drafting an AI policy is insufficient. Effective implementation requires robust policy training, a clear system of oversight or administration, and a commitment to continuous dialogue with both internal staff and external stakeholders, including other news outlets and the audience itself. These interviews, edited and condensed for clarity, provide a comprehensive roadmap for navigating the complexities of AI integration in journalism.
The New Technological Paradigm: Why AI is Different
The current wave of AI innovation, particularly the widespread accessibility of generative AI tools like ChatGPT since late 2022, represents a qualitative leap from previous technological disruptions in journalism. While the internet era introduced new distribution channels and social media transformed audience engagement, AI directly impacts the production of journalistic content. This fundamental shift means that AI can directly influence the accuracy, authenticity, and ethical integrity of published work. The potential for AI "hallucinations"—fabrications presented as facts—poses an unprecedented risk to journalistic credibility.
The speed of AI development further distinguishes this era. Algorithms are constantly evolving, capabilities are rapidly expanding, and new tools emerge almost daily. This relentless pace renders static policies quickly obsolete, demanding a dynamic and adaptable approach from news organizations. Unlike a new content management system or social media platform, AI is not a fixed tool but a continually evolving ecosystem.
Crafting Foundational Policies: Values and Resources
Before a single word of an AI policy is written, Anika Collier Navaroli advises organizations to first define what AI use means in the context of their core values. If transparency, for instance, is central to an organization’s operations, then this principle must be at the heart of its AI framework. This values-first approach ensures that technological adoption remains aligned with journalistic ethics rather than being driven solely by efficiency or novelty.
Several newsrooms have found valuable guidance in existing resources. The "AI Ethics Starter Kit" from Poynter is frequently cited as a crucial tool for building policies and processes. Other organizations have drawn inspiration from the guidelines published by the Associated Press, the Society for Professional Journalists (SPJ), and the American Journalism Project. These frameworks offer a starting point for grappling with the complex ethical considerations inherent in AI adoption.
A critical step involves cataloging current and prospective AI applications within the newsroom. A 2025 survey by the Institute for Nonprofit News (INN) identified data analysis, grant application drafting, and data scraping as common AI uses among its members. More routine applications include interview transcription, grammatical suggestions, and document review. Significantly, most newsrooms with established AI policies explicitly prohibit the use of generative AI for writing entire stories or creating images without direct human involvement and robust editorial oversight. This distinction between AI as an assistive tool and AI as an autonomous content creator is a cornerstone of current policy.
Laura Zelenko, Global Head of Editorial Standards at Bloomberg News, articulated a set of core principles that underscore this human-centric approach: "We have emphasized that nothing replaces original reporting. And when we discuss AI, we focus on responsible use of AI tools in a newsroom to help analyze big datasets, speed up research efforts, and automate more routine tasks—with a core principle being that we ensure a human is always involved and nothing gets published without human oversight." She further elaborated on key tenets: "Other core principles are transparency (clearly label when AI is being used in a significant way), accountability (every journalist is responsible for every word they publish, and plagiarism is a fireable offense), and authenticity (don’t use AI to write or edit a story from scratch)."
Tess Jeffers of the Wall Street Journal highlighted the iterative nature of policy development. "At the time of our first policy draft, in 2023, very few AI guidelines had been published. Wired was one of the first, and we were motivated by their work," she explained. Research by Hannes Cools and Nicholos Diakopoulos provided a valuable "roundup" of emerging policy "flavors" across the industry. Over three years, the Journal has continually refined its policies, maintaining a running list of what works for different publishers and what aligns with their own needs. Jeffers emphasized the careful balance struck between encouraging experimentation and safeguarding journalistic integrity and intellectual property. "It was very important for us to lead with excitement rather than fear, which is why our guidelines open with the line ‘Advances in artificial intelligence create incredible opportunities for our newsroom.’ And only when you get five grafs down do you see the first ‘don’t do this’ statement."
Felicitas Carrique, Executive Director of the News Product Alliance, stressed that policy must be audience-centric: "Before the policy, before the tools, organizations need to be able to clearly answer: Who are we serving, and what do they actually need from us now, and in the future? Policy is only as good as the thing it’s serving, and the starting point comes from answering those questions clearly."
Understanding the Scope: AI’s Practical Applications and Limitations
The practical application of AI in newsrooms extends beyond theoretical discussions. Eileen O’Reilly, Head of Standards and AI Practices for Axios, outlined their approach: "We want our policies to be editorially driven and our tools also to be editorially driven. Our policy right now is that we do not create any text using generative AI but may use it to help create some data visualizations." She noted the use of "suggesting capacities, like suggesting a stronger headline or alternative text for images," always with human editorial oversight. O’Reilly also envisioned AI filling critical gaps in local news: "My hope is that AI can help us scale into areas where there are already news deserts. There aren’t enough people covering city council and school board meetings already… I would love it if AI could help us get transcriptions and summarize them so reporters can see if there’s anything important we should cover. Then the reporter can follow through, call people, and do their investigative reporting based on what they find from these transcript summaries. To me, this is where AI would be golden."
Tav Klitgaard, Group CEO and cofounder of Denmark’s Zetland, advocated for a principles-based approach over rigid rules. "In developing a policy: don’t overthink, and don’t be too concrete. We like working with principles instead of rules. Rules, model recommendations, and dos and don’ts all quickly become outdated. Principles should be able to last longer." Zetland‘s core vision is to feel like a "very human product," presenting the challenge of using AI to become more human, not less.
Cynthia Tu, a Data Reporter and News Technology Specialist at Minnesota’s Sahan Journal, provided a concrete example of collaborative policy drafting. She and her manager, Chief Growth Officer Michael Tortorello, formed a "draft committee," engaging department heads, managers, and the entire staff through one-on-one meetings, breakout sessions, and open office hours to gather feedback on tools, use cases, concerns, "hopes and dreams." This inclusive process ensured that the final policy reflected the diverse perspectives and needs of the newsroom.
Beyond the Document: Training and Internal Adoption
A well-crafted policy document is merely the foundation; its efficacy hinges on how it is implemented and integrated into daily operations. Without robust support and continuous training for staff and contributors, policies risk becoming inert PDFs, consulted only during orientation and quickly forgotten. To transform policy into practice, news organizations must invest in ongoing education, foster a culture of experimentation and sharing, and provide platforms for collective learning.
A recent FT Strategies report on newsroom routines identified significant barriers to AI adoption, primarily human-centric rather than technical: "the top three barriers to AI adoption are skills gaps (61 percent), cultural resistance and skepticism (52 percent), and unclear use cases (45 percent)." Navaroli underscored that support and training extend beyond mere compliance. "This is not just ‘dos and don’ts.’ This is an ongoing conversation that you’re having with your colleagues, whom you respect, about emerging technology issues that are going to be challenging and difficult."
Jane Barrett of Reuters detailed their comprehensive training strategy: "We require everybody to take a refreshed AI 101 training every year so they are up to date with our policies, changes in the AI world and the tools at their disposal, how some of their peers are using AI in their journalism. We also have monthly town halls, regular demos, and emails to keep people engaged, and provide space for sharing and debate."
Tav Klitgaard shared Zetland‘s experience with building an in-house transcription service, Good Tape, and recruiting AI experts to work alongside journalists and tech teams. This created an early "positive buzz" around AI. "We formed a group of internal champions, and they shared their knowledge and also arranged informal sessions for the company. The most important thing has been to share best practices. The policy is not at all important. I doubt anyone ever looks at it… But the editorial ethics and culture are quite ingrained. What we do is to spend time talking about data privacy and what the ingrained rules of journalism mean in this new reality."
Martin Schori, former Director of AI and Innovation at Sweden’s Aftonbladet, emphasized a progressive training approach: "We started with training. We trained all our journalists in prompting, and then it was ChatGPT—the only thing that was around back then. Then we moved on to editorial tools, transcription tools, summarizations, SEO stuff, text-to-speech."
Tess Jeffers of the Wall Street Journal further elaborated on effective training: "While it’s important to have a policy document, we’ve found it even more useful to discuss those guidelines in person and explain things like why using enterprise-grade tools is critical to protect our journalism and IP. ‘AI 101’ training sessions have been a huge help in explaining the ‘why’ behind our rules, rather than just handing out a list of requirements." The Journal also leverages "show-and-tell sessions" where colleagues demonstrate AI use for reporting, fact-finding, editing, or "vibe coding" new tools, fostering practical application and peer learning.
Establishing Robust Oversight and Governance
Even with the most meticulously crafted policies and comprehensive training, newsroom leaders must anticipate that errors will occur. Staff, and indeed AI itself, will make mistakes; "hallucination is inevitable." There may also be instances where AI proves more burdensome than beneficial, prompting internal and external questions about its value. To navigate these challenges effectively, a proactive governance structure—such as a task force or a regularly convening oversight group—is indispensable. This ensures preparedness and adaptability.
The FT Strategies report indicates that newsrooms demonstrate significantly higher confidence in their technology choices when journalists, not just senior management, are actively involved in the decision-making process. Navaroli recommends that advisory groups be diverse, including standards professionals, technical experts, and representatives from various departments like editorial and visuals. She also suggests term limits to prevent undue influence: "It’s important that you rotate folks, so that they don’t get too much power." Regular meetings are crucial for identifying policy gaps, addressing unexpected use cases, and determining when updates are needed.
Tess Jeffers highlighted the importance of inclusive governance at the Wall Street Journal: "Building our AI guidelines was a top priority for our newsroom AI task force immediately after we formed the group. The task force includes both AI enthusiasts and skeptics from across our photo, audio, graphics, audience, data journalism, newswires, coverage, and standards teams." Their guidelines are treated as a "living document," reviewed every six months or after significant technological shifts.
Matthew G. Miller, Senior Executive Editor at Bloomberg News, described their "AI Advisory Board" as a "cross-functional group of leaders from our editorial, research, and product divisions, such as breaking news, television, data journalism, Bloomberg Economics, and Bloomberg Intelligence." This global team acts as advocates and champions for technology while ensuring adherence to editorial guidelines.
Matthew DeFour, State Bureau Chief at Wisconsin Watch, noted that a highly publicized incident at the Wisconsin State Journal—where a journalist was fired after an AI-hallucinated business owner and quote appeared in a story—spurred significant interest in forming their own committee. This committee, comprising representatives from business, editors, and reporters, is now actively developing Wisconsin Watch‘s AI policy, ensuring equal representation from both union and management.
Sashka Koloff from the ABC offered a compelling case study on ethical deliberation. When producers considered using AI to recreate the voice of a deceased convicted pedophile, Rolf Harris, for a documentary series, the internal ethical framework was rigorously applied. Questions arose: "What is our level of comfort? How do we do this ethically? How do we tell the audience this? Is there any way that the audience could be misled or confused by what they’re seeing, and how would that affect trust?" Ultimately, the producers opted against AI voice recreation, finding a more authentic storytelling approach using real recordings. This example highlights the critical role of ethical gatekeeping and audience perception in AI implementation.
Jane Barrett of Reuters underscored the need for continuous vigilance and the "kill switch" mechanism. "The AI world moves incredibly fast, so all of us publishers need to keep reviewing our policies. For Reuters, that means having a senior editor responsible for reviewing our guidance regularly and an AI Governance Committee, chaired by our editor in chief, which meets monthly to look at the tools and solutions people are using and building, testing them against our standards, feeding back before approval, and always maintaining the ability to hit the kill switch if something goes awry."
Eileen O’Reilly of Axios also stressed the importance of ongoing testing, citing their internal tool, Axiomizer, which helps reporters refine copy and ensure optimal usage of axioms. While useful for grammar and style, she noted its current limitations for fact-checking. "We keep running tests through it, and it won’t find everything, or sometimes it will find something that is correct and flag it as incorrect. We’re getting much closer to a good fact-checking tool."
The Imperative of Transparency and Collaboration
In a landscape where all news organizations face similar, rapidly evolving challenges with AI, an open and collaborative approach is not merely beneficial but essential. Navaroli distinguished between two key audiences for AI policies: "There are often two different audiences that you want to think about. There is the public-facing, and an external policy that you’re going to publish on your website, an AI policy or a data privacy policy. And then you have internal rules: the things that you keep internally for your employees that go into greater detail."
Sharing best practices, case studies, and even "counterexamples" with peers is vital. Conferences, informal discussions among standards editors, and dedicated forums provide invaluable opportunities to discuss trade-offs, assess risks, and learn from mistakes. Felicitas Carrique articulated this necessity: "The field is moving faster than any single organization can keep up with alone, and sharing, which journalism has historically been bad at, is now a competitive necessity. The organizations that try to figure this out in isolation are leaving hard-won learning on the table."
Sashka Koloff reiterated the need for continuous review, emphasizing alignment with audience expectations. "You can’t just set some policies and forget. You need to be in line with audience expectations, making sure that they actually work in practice… It’s foolish to think that you set forward policies and then not review them. We have to be in constant review." Jane Barrett observed an interesting shift at Reuters: "One of the really interesting changes we are seeing is that some of our journalists are becoming smart product builders and some of our engineers are spotting journalism problems to solve."
Martin Schori highlighted a crucial audience insight: "Most people are not super interested in exactly what tools you’re using for various tasks. I think the only thing they want to know is: Is it a human or a journalist who makes the bigger decisions? Who chooses what you are reporting, for example, and who’s responsible for it? People are concerned about that." This underscores that transparency about human oversight and accountability is often more important to audiences than the specific AI technologies employed.
Cynthia Tu of Sahan Journal exemplified this transparency, particularly for niche audiences. As a nonprofit serving immigrants and communities of color in Minnesota, Sahan Journal is acutely aware of potential AI biases and disparities. Tu shared how she transparently communicated their use of AI tools in reporting and business operations in their weekly newsletter, detailing challenges, concerns, and decisions. This proactive communication led to positive feedback and increased accessibility, whereas merely mentioning AI without context often elicited negative reactions. "And that’s why our AI policy is so important," she concluded.
Conclusion: Adapting to the Future of Journalism
The integration of artificial intelligence into newsrooms is not a temporary trend but a fundamental, ongoing transformation. The experiences and insights from leading news organizations underscore a multi-faceted approach to navigating this new frontier. Key takeaways include the imperative of values-driven policy development, the necessity of continuous staff training and a culture of experimentation, the establishment of robust and inclusive governance structures, and an unwavering commitment to transparency with both internal teams and the public.
Ultimately, AI is poised to serve as a powerful assistant, enhancing journalistic efficiency, expanding coverage into underserved areas, and freeing up human journalists to focus on the irreplaceable core of their profession: critical thinking, investigative reporting, empathetic storytelling, and ethical judgment. The journey of adapting to AI is an iterative one, characterized by constant review, learning from mistakes, and collaboration across the industry. By embracing these principles, news organizations can harness the transformative potential of AI while safeguarding the trust and credibility that are the bedrock of journalism.
