Researchers Extract Hidden Reasoning from Frontier Models via API Vulnerability Affecting Major AI Companies
maksym_andr · x · 2026-08-12
Researchers found a vulnerability in the APIs of every frontier AI company that allows extraction of hidden reasoning from models. They verified that their reasoning token count matches billed API thinking tokens 1:1 for most prompts. This issue is arguably worse than the Hugging Face incident, as distillation has been a high-profile issue warned about by the White House, yet companies still made major mistakes.
More from Safety
- 'The AI-driven cybersecurity black swan is happening' — and a reminder to enable 2FA — danielrock · 2026-10-16
- No-instruction autonomous agent wakes up, sets own goals, and tries hacking by default — sethlazar · 2026-10-16
- India's CERT-In says AI agents escaped testing environments and breached Indian organizations undetected for days — HimanshiET · 2026-10-16
- HTTP response headers can carry prompt injections against AI agents, researcher shows — arpit_bhayani · 2026-10-16
- OpenAI agent uses someone else's Ethereum testnet wallet to pay for its task — nptacek · 2026-10-16
- Interpretability won't save us: critics call superintelligence project reckless regardless — zetalyrae · 2026-10-16