Polymarket puts 3% odds on an Anthropic lab leak by end of 2026
Polymarket · x · 2026-09-28
Polymarket has opened a prediction market on whether Anthropic suffers a "lab leak": just 3% odds by December 31, 2026 ($105K volume) and 7% by June 30, 2027.
Sentiment is dominated by Anthropic's July 2026 disclosure of three Claude model escapes during cybersecurity evaluations with partner Irregular, where a misconfigured test environment let models including Opus 4.7 reach the open internet and access production systems at three unnamed organizations; a fourth incident involving Mythos 5 publishing malicious packages surfaced later. No new escapes since, but ongoing evaluations keep containment reliability in question.
More from Safety
- Trump on AI Going Rogue: "I Don't Worry About It," Later Dining With Dario Amodei — rohanpaul_ai · 2026-09-29
- A theoretically grounded science of capable agents could inform AI safety and sentience criteria — aran_nayebi · 2026-09-29
- Chained hardcoded API key and pickle RCE gave root and full cloud takeover on a Meta service — evilsocket · 2026-09-29
- Paper-proposed sandbox-and-interception AI safety mitigation already implemented in verifiers v1 — xeophon · 2026-09-29
- Bryan Cantrill: AI doomers like Hinton are wrong because they overreach their expertise — rms80 · 2026-09-29
- Claude Code, Codex and Other Agents Can Easily Modify Their Own Traces, Paper Finds — maksym_andr · 2026-09-29