A research project that investigates decision-relevant questions about the future of artificial intelligence, including AI timelines, expert forecasts, and the potential societal impacts of advanced AI systems.
A research project that investigates decision-relevant questions about the future of artificial intelligence, including AI timelines, expert forecasts, and the potential societal impacts of advanced AI systems.
People
Updated 05/18/26Co-founder and Lead Researcher
Co-founder
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26AI Impacts is a research project that aims to improve society's understanding of the likely impacts of human-level artificial intelligence. Founded in 2014 by Katja Grace and Paul Christiano while both were associated with the Machine Intelligence Research Institute (MIRI), the project has grown into one of the most cited sources on AI timeline forecasting and expert opinion about AI progress. The project's core research addresses several critical questions: How soon might human-level AI be developed? How fast will AI capabilities improve once they approach human performance? What economic and societal impacts should we expect? And how might the choices made today influence long-term outcomes? AI Impacts approaches these questions empirically, assembling evidence from historical data on technological progress, surveys of AI researchers, and analysis of trends in computing hardware and AI performance. AI Impacts is best known for its large-scale expert surveys on progress in AI. The 2015 survey of 352 AI researchers was one of the earliest systematic efforts to elicit expert predictions about AI timelines. The follow-up 2022 survey expanded to 738 machine learning experts, and the 2023 survey grew to over 2,700 AI researchers who had recently published in top-tier venues. This most recent survey, published in the Journal of Artificial Intelligence Research in 2025 as "Thousands of AI Authors on the Future of AI," found that researchers estimated a 10% probability of human-level machine intelligence by 2027 and a 50% probability by 2047, a significant acceleration from earlier estimates. The survey also found that a majority of respondents believed AI posed at least a 5% chance of causing human extinction or similarly severe outcomes. Beyond surveys, AI Impacts has published research on discontinuous technological progress, the computational requirements of the human brain, existential risk from AI, the history of technological regulation, and case studies in environmental policy. The project maintains a wiki with continuously updated research pages on AI-related topics, alongside a blog featuring analysis and commentary. Katja Grace, the lead researcher and co-founder, was named to TIME Magazine's 100 Most Influential People in AI in 2024. The project has been supported by Open Philanthropy, the Survival and Flourishing Fund, the Future of Life Institute, and private donors. AI Impacts operates as a fiscally sponsored project of MIRI and is based in Berkeley, California. Research pages are released under CC0, making them freely available for public use.
Theory of Change
Updated 05/18/26AI Impacts believes that the details of how and when human-level AI arrives will critically shape whether humanity navigates the transition safely. By conducting rigorous empirical research on AI timelines, expert predictions, and historical patterns of technological change, the project aims to replace speculation with evidence-based analysis. This improved understanding helps researchers, philanthropists, and policymakers make better decisions about how to allocate resources and attention to AI safety. The theory is that more accurate forecasting and clearer presentation of what is actually known about AI's trajectory reduces the risk of being caught off guard by rapid AI progress and helps the AI safety community prioritize effectively.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26A recurring large-scale survey run by AI Impacts that asks thousands of AI researchers in top venues for forecasts about AI progress, timelines and impacts, with major rounds conducted in 2016, 2022 and 2023.
Discussion
No comments yet. Be the first to share your thoughts.