The scoreboard for AI forecasting
CEOs, researchers, Wall Street, the press — the AI Prediction Index tracks their public predictions, puts a clock on every deadline, and scores each call against what actually happened.
Predictions tracked
20
and growing daily
Resolved
10
deadline passed
On the clock
10
counting down
Hit rate of resolved
20%
partial credit = ½
10 predictions whose deadline has passed
| Group | Tracked | Resolved | Hit rate |
|---|---|---|---|
| Tech CEO | 9 | 5 | 30% |
| Researcher | 5 | 4 | 13% |
| Thought Leader | 3 | 1 | 0% |
| Economist | 1 | 0 | pending |
| Wall Street | 1 | 0 | pending |
| Investor | 1 | 0 | pending |
“AI will pass a valid Turing test, achieving human-level intelligence, by 2029.”
Conversational tests are increasingly passed, but Kurzweil's own definition requires expert-level performance across the full range of human cognition.
“People should stop training radiologists now. It's just completely obvious that within five years deep learning is going to do better than radiologists.”
By 2021 AI had become a useful assistive tool in radiology, but radiologist demand grew — the US faced a radiologist shortage. Hinton later acknowledged the timeline was wrong.
“Tesla will have one million robotaxis on the road next year (2020).”
Zero Tesla robotaxis operated in 2020. Tesla's first limited, supervised robotaxi pilot did not launch until 2025 — with a small fleet, not a million.
“Powerful AI — smarter than a Nobel Prize winner across most relevant fields — could come as early as 2026.”
Framed explicitly as an 'as early as' estimate, not a commitment.