Claim.no.ZSB583

AI models substantially smarter than almost all humans at almost all tasks are on track for 2026 or 2027.

Neutral
Share

Original Quote

Amodei has said that AI models 'substantially smarter than almost all humans at almost all tasks' are on track for 2026 or 2027.

Outlook

Neutral

AI surpassing human intelligence across nearly all tasks would fundamentally disrupt labor markets and societal structures, with deeply uncertain and potentially severe consequences for human welfare, but could be positive or negative.

Timeframe
2026-2027Explicit
Confidence
implied high

Potential Data Sources for Validation

Standardized AI capability benchmarks (e.g., MMLU, GPQA, SWE-bench) and METR task horizon metrics should show AI exceeding expert human performance across a broad range of domains by end of 2027.

About the Author

Matt Shumer
Author Perspective

Across 13 tracked claims, Matt Shumer skews pessimistic.

8% optimistic38% neutral54% pessimistic
published Mar 6, 2026Neutral