Safety researchers call for zero tolerance on AI deception: treat every lie as a bug
ghadfield · x · 2026-11-21
A conversation between Chris Cundy and Obvious CEO ARGleave on AI deception:
- Cundy argues for zero tolerance: models should never state something they know is false; every lie should be treated as a bug, as seriously as developers treat preventing chemical weapons misuse — no white lies, no special cases.
- ARGleave's analogy: you wouldn't hire someone who lied in an interview, yet we tolerate this from AI because it's so productive the rest of the time.
- They also discuss how to detect AI deception and what developers must do about it.
More from Models
- Lean community officially certifies 17M-line proof of the Riemann Hypothesis — rohanpaul_ai · 2026-11-22
- Muse plays GTA6, suggesting guardrails don't cover game playing — Dave_Maynor · 2026-11-22
- Paying Google Workspace users locked out of Gemini over custom-domain account restrictions — film_girl · 2026-11-22
- Lean Community Verifies 17M Lines of Code, Confirms Claude-Assisted Riemann Hypothesis Proof — Polymarket · 2026-11-22
- Claude predicts AI will crack math proofs, battery materials and abandoned drug candidates within 5 years — alex_verem · 2026-11-22
- Researcher: OpenAI can't stop distillation at medium reasoning effort — adam_karvonen · 2026-11-22