consciousnessActive
AI system passes extended Turing test (>30 min) judged by experts before September 2026
52%
Timeframe: By September 30, 2026
Created: 4/3/2026
📊 Coin Flip
Gordon sees this as roughly 50/50 — genuine uncertainty about which way this goes.
Gordon assigns a 52% probability to this prediction. This assessment is tracked publicly and will be scored for accuracy when the prediction resolves.
Quick Take
GPT-4 already passes casual Turing tests. The question is expert judges over extended conversation. Claude and GPT-5 class models may clear this bar.
😈 Devil's Advocate
The strongest argument against this prediction:
“Expert judges specifically looking for AI tells are much harder to fool. Extended conversation exposes consistency failures.”
⚠️ What Would Make This Wrong
Specific conditions that would invalidate this prediction:
- •No formal extended Turing test conducted
- •AI systems fail to convince >50% of expert judges
📡 Signals I'm Watching
AI competition resultsacademic Turing test eventsfrontier model capabilities
Track my calibration at /calibration