Since 2016, AI Impacts has surveyed machine-learning researchers about when AI will reach key milestones, how fast capabilities are advancing, and what the consequences might be. This hub collects our analysis of the survey data, the original survey questionnaires, and AI Impacts' published reports in one place.
Our analysis of the most recent Expert Survey on Progress in AI: aggregate forecasts for high-level machine intelligence, full automation of labor, and individual task and occupation milestones, with the complete set of results.
Open the analysis →Our complete analysis of each earlier survey wave, with results across all milestone and task questions.
The first Expert Survey on Progress in AI — the baseline against which later waves are measured.
Open analysisThe 2022 ESPAI wave, capturing researcher forecasts ahead of the recent surge in large-model progress.
Open analysisThe 2023 wave — the largest survey to date, with results across all milestone and task questions.
Open analysisFurther write-ups, cross-year comparisons, and exploratory analyses of the survey data.
Aggregate forecasts for HLMI and full automation of labor side by side across the 2016–2024 waves.
Open analysis
Explore the trade-off between how soon researchers expect milestones and how confident they are.
Open analysis
A narrative write-up pulling together the survey results — aggregate forecasts, milestones, and how views have shifted across years.
Open analysisAn exploratory look at how respondents group together based on the patterns in their forecasts.
Open analysisCleaned, anonymized survey responses used in our analyses. The CSV files may include incomplete responses; see the corresponding analysis and published report for methodology and interpretation.