Tarbell aims to distribute at least $10M in funding by the end of 2027 to support ambitious journalism on AI.

We understand that an outlet, newsroom department, or series focused on AI may span a few areas. We will look to fund outlets, departments, and series that predominantly cover one or more of the following topics:

Coverage that seriously engages with the possibility that AI matching or surpassing human capabilities across most cognitive tasks could be developed in the coming years, and what this would mean for society. This is a contested and uncertain topic — which makes high-quality journalism about it especially valuable. We support diverse perspectives here, including work that challenges our own assumptions or presents evidence that cuts against prevailing narratives in either direction.

Explanatory journalism and analysis that educates and informs key decision-makers about AI development and deployment. This includes reporting on factors driving AI progress (such as the semiconductor supply chain, AI adoption rates, and export controls), current AI capabilities, and technical explainers of important research developments. We believe that increasing the density of accurate and reliable information in the public media on these topics leads to better-informed decisions by policymakers, companies, and civil society.

Accountability reporting on voluntary safety commitments, internal practices, and corporate behavior at leading AI companies (OpenAI, Anthropic, Google DeepMind, Meta, xAI, and others). Without binding government regulation in most jurisdictions, society currently relies on these companies following through on voluntary safety commitments. Journalism can play a key role in ensuring these commitments are upheld — and in surfacing cases where they aren’t, such as companies releasing models that don’t meet their own stated safety thresholds, or failing to follow through on resource commitments to safety teams.

Investigations into how AI companies and other relevant actors influence policy in key jurisdictions, including at the federal and state level in the US, as well as in other influential jurisdictions like the UK and EU. Lobbying and campaign spending significantly shape policy outcomes in other domains, and early evidence suggests AI is no exception. Investigative reporting can help ensure that AI governance is shaped by public interest considerations rather than by which actors spend the most on lobbying.

Reporting that deeply engages with proposed AI policies in the US at the federal level and in key states, the potential impacts these policies could have on innovation and safety, and the surrounding dynamics shaping them. While we prioritize the US, we will also consider policy reporting proposals in other jurisdictions that are influential on AI policy, such as the UK and the EU. We believe high-quality AI policy coverage should reflect a wide range of political perspectives, including perspectives that are currently underrepresented. Sustained journalistic engagement with the policy process can improve outcomes — for example, EU policy reporters’ multi-year coverage of the continent’s AI Act is widely seen as having improved the final legislation.

China is likely the second most important jurisdiction in AI development, but Western media coverage often lags behind or focuses narrowly on US-China competition rather than substantive developments. Important coverage gaps include: internal perspectives from Chinese AI and security institutions, the impacts of AI diffusion in Chinese society, technical advancements at Chinese AI companies and their global implications, and the state of US-China dialogue on AI safety and governance.

Reporting that surfaces real-world harms or near-harms from deployed AI systems, especially those related to catastrophic risks or major societal disruptions. Many AI incidents — such as the OpenAI-Hugging Face hack — are poorly understood even by employees at frontier AI companies, who often first learn about the impacts through media reporting. A strong cadre of reporters covering this beat could uncover early evidence of societal impacts from AI, helping decision-makers identify and respond to warning signs before they escalate.

Reporting that generates new evidence on how AI is — or isn’t — reshaping work. As AI capabilities expand, there is intense debate but limited rigorous evidence about the actual impact on employment, wages, and the nature of work. We’re particularly interested in reporting that goes beyond speculation and anecdote to provide data-driven analysis of labor market changes as they unfold.

NPR’s Power and Influence team reports on how technology intersects with power, money, politics and influence. Bringing decades of combined experience, this team leads NPR’s coverage of frontier AI across radio, podcast, and digital platforms. Tarbell has awarded the team a grant of $250,000 to support its coverage.

“AI is an increasingly central storyline on many of the beats we’re covering at NPR. We’re working hard to inform our audience how the technology works, the ways in which it’s affecting their lives and the ways in which it’s remaking the world. We’re thrilled to have Tarbell’s support to help us cover AI with even more ambition and depth.”

– Brett Neely, Supervising Senior Editor, NPR

Machine Gods is a new video-first show from Casey Newton and Kevin Roose, the creators of the hugely successful New York Times podcast Hard Fork, and two of the most recognized names in AI journalism today. The twice-weekly show, launching in partnership with NPR in October, will combine sharp news, expert analysis, and high-impact interviews with the builders and thinkers shaping the AI future. Applying journalistic rigor to both the promise and perils of the technology, Machine Gods will bring independent reporting on frontier AI companies directly to millions of viewers and listeners on YouTube, major podcast platforms, and public radio station feeds—completely free and non-paywalled. Tarbell is proud to support their launch with a grant of $480,000.

“We are longtime admirers of Tarbell’s mission of funding high-quality AI journalism, and we’re thrilled to be a Tarbell grantee. Tarbell’s support will help us scale our show quickly, experiment with new formats, and bring more grounded, independent AI reporting into the world at a time when it’s critically needed.”

– Kevin Roose, Co-host, Machine Gods