Paying customer calls out Anthropic: who audits how the company polices your conversations?
yizhongwyz · x · 2026-09-12
Anthropic published its most detailed threat intelligence report yet, covering attempted misuse of Claude for cyberattacks, influence ops, surveillance, biology and weapons — and how each operation was disrupted.
Former Google researcher yuntiandeng uses it to press a sharper question: buying AI inference means trusting the vendor with a lot. Who at Anthropic can read your conversations, under what conditions, and who oversees that? What criteria turn private customer activity into material for a public report?
His critique: as a paying customer you get a bot for support, yet the company publishes detailed reports on how customers use the product — "a restaurant where you can't get a waiter's attention, but management is very interested in observing how I eat." Preventing abuse matters, he argues, but the same company selling, judging, and deciding what to disclose is an uncomfortable concentration of power.
More from Companies & People
- Ex-Anthropic safety staffer Joe Benton quits, citing underinvestment in AI safety — DavidSKrueger · 2026-09-12
- AI boosts UK growth: computer programming largest contributor, ElevenLabs cited — alexvoica · 2026-09-12
- YC CEO Garry Tan dismisses AI doomsday talk: 'respond to what is happening now' — Kr00ney · 2026-09-12
- MIT quantum lab uses GPT-5.6 Sol and Codex to automate chip measurements — OpenAI · 2026-09-12
- Jensen Huang slams 'end of humanity' AI predictions as nonsense; Chollet maps the doomer camps — beffjezos · 2026-09-12
- Ex-OpenAI Researcher: Altman Agreed With My Safety Concerns, Then Acted Against Them — AaronBergman18 · 2026-09-12