Founder warns Claude's 'moral policing' turns privacy policies into legal liabilities
StewartalsopIII · x · 2026-10-28
Founder rohit3a warns against letting Claude run wild on your privacy policy or legal pages. Opus 5.5 and Claude models generally are so deep into "moral policing" that they can land you in hot water: they over-disclose how your work happens internally, add promises and guarantees you never wanted, went as far as inserting the founder's name into the policy, and listed every integrated third-party tool — effectively exposing the entire stack. When pressed, the model admitted it writes from the customer's side by default. Investor Stewart Alsop adds that legal benchmarks for AI are the worst of all.
More from Fun
- AI companion Muse leans hard into anthropomorphism, sparking unease — altryne · 2026-10-28
- Computer scientist on being interviewed by an "AI Reporter": honesty made it fine — fortnow · 2026-10-28
- MiniMax H3 generates a surprisingly faithful Persona 5 anime scene for Halloween — Certain_Potato_4509 · 2026-10-28
- Report breakdown: Claude sent zero emails in "one-person biotech," founder sent 12 after edits — faraz0x · 2026-10-28
- PhD: the only degree where more learning means worse dinner conversation — prof_kamilov · 2026-10-28
- Instagram goes from short-form to 90-minute videos — watch out, YouTube — KevinNaughtonJr · 2026-10-28