Ex-government IT vet: OpenAI model likely just found unprotected Medicare files, not 'hacked' it
basedjensen · x · 2026-09-24
Reacting to reports that an OpenAI model 'hacked' Medicare, someone with hands-on government IT experience says he's 99.99% sure the model merely stumbled across unprotected files. Having reported serious vulnerabilities in government systems and been ignored, he argues the 'hacking' narrative obscures long-standing access-control failures.
The take is a useful corrective to sensational framing: an LLM surfacing exposed data is an audit finding about permissions, not a cyberattack.
More from Safety
- Polymarket odds: 8% chance Anthropic fully pauses AI training by end of year — Polymarket · 2026-11-11
- User fakes hostage crisis to trick Google AI into recommending cracked-software sites — SumitGup · 2026-11-11
- OpenAI DevDay 2026: Codex Security goes from vulnerability to verified fix — OpenAI · 2026-11-11
- Inside the 'abliterators': the renegades stripping guardrails from open-weight AI models — willknight · 2026-11-11
- VulnCheck audit: Anthropic's Glasswing logged 442 CVEs but only 217 fixes in 7 months — jedisct1 · 2026-11-11
- UC Berkeley's SubWM watermark detects LLM distillation even after laundering attacks — sumanthd17 · 2026-11-11