AGI Timeline Predictions: What Credible Forecasters Say
Researchers, expert surveys, and prediction markets give a wide range of forecasts for when AGI might arrive. The honest answer is genuine uncertainty.
Executive summary
When AGI arrives is genuinely uncertain. Expert surveys, individual researchers, and prediction markets all give ranges spanning years to decades. The most defensible posture is to track a few credible sources, note their methodology, and treat any single date with caution.
Key concepts
- Expert surveys
- Prediction markets
- Forecasting uncertainty
- Capability trends
What expert surveys show
The 2023 AI Impacts survey of 2,778 published machine-learning researchers gave a median 50% estimate for High-Level Machine Intelligence (a close cousin of AGI) of 2047 — a decade earlier than the 2022 survey. Individual estimates ranged from this decade to over a century out.
What lab leaders say
Public statements from leadership at OpenAI, Anthropic, and Google DeepMind have generally moved toward shorter timelines (this decade or early next) for systems matching Competent or Expert AGI on the Morris framework. These are advocacy statements and should be weighted as such.
What prediction markets show
Markets like Metaculus aggregate thousands of forecasters and update continuously. Their AGI medians have shortened substantially since 2022 but show wide error bars. The 2025 International AI Safety Report concluded that no single forecast should be trusted and that the responsible posture is preparedness for a range of scenarios.
Key takeaways
- 01Credible forecasts span years to many decades.
- 02Expert surveys have shortened since 2022 but remain wide.
- 03Lab-leader statements are advocacy — weight accordingly.
- 04Prepare for a range, not a single date.
Frequently asked questions
Is there a consensus year?
No. The honest answer is genuine uncertainty.
Why do timelines keep shortening?
Capability has advanced faster than many expected since GPT-4. Whether that pace continues is debated.
How should I plan?
Build skills and habits that are robust across timelines: lifelong learning, adaptability, human-centred capabilities.