Aidan Clark says safety techniques are easier to undo once they’re known
_aidan_clark_ · x · 2026-07-23
Aidan Clark argues that if people know a technique, it becomes easier to undo — so sharing safety techniques may weaken them.
He also says the team would be open to partnerships that help harden external models, including open-source ones, but they would be much less eager to share the techniques themselves.
Related event: GPT-OSS Security Transparency Debate: Secrecy vs. Openness(7 posts)→
More from Safety
- Anthropic's S-1 unusually discusses AI sentience and moral status risks — birchlse · 2026-10-27
- Researcher Asks Whether Suspicious Model Behavior Can Ever Be Explained — or Faked — jd_pressman · 2026-10-27
- UK AISI papers: no good safeguards yet against misuse of biological AI tools — davidmanheim · 2026-10-27
- Washington's message to AI execs: Please shut up — TobyWalsh · 2026-10-27
- Are Meta Muse, Grok Bot and OpenAI Dots really safe hands for our privacy? — Neat-Flight8381 · 2026-10-27
- PHP2Shell exploits confirmed in the wild, attackers may use agentic AI for post-exploitation — AravSrinivas · 2026-10-27