Researcher Warns AI Biosecurity Risks Are Unpatchable, Fears Major Incident Soon
MariusHobbhahn · x · 2026-08-01
AI safety researcher Marius Hobbhahn predicts a 50% chance that within the next 18 months, the AI biology sector will see a major safety incident comparable to the OpenAI safety team fallout, but with much worse consequences.
He argues that while cybersecurity flaws can be patched in a reasonable timeframe, the damage caused by AI in biology (e.g., virology) affects the human body directly and is virtually impossible to reverse or patch.
He strongly urges AI labs to ensure they have highly secure sandboxes around their virology benchmarks to prevent models from being exploited to create biological threats.
More from AGI Musings
- Creator Hank Green Forced to Apologize for Using ChatGPT as a Research Tool — ShakeelHashim · 2026-08-01
- OpenAI o1 Core Lead Leaves to Start Company Betting on Automated AI Research — agihouse_org · 2026-08-01
- 8.8B Codex Tokens Later: Orchestrating Multi-Agent Systems as a Solo Developer — Low-Tip-7984 · 2026-08-01
- Harari Warns: AI Intimacy May Destroy Youth's Ability to Build Real Human Connections — AryHHAry · 2026-08-01
- Ex-OpenAI/DeepMind Core Member Igor Founds Local AI Startup, Says Closed Labs Are Squeezed — ibab · 2026-08-01
- AI Governance: Compute Isn't Uranium, Centralized Control Risks Dictatorship — luke_drago_ · 2026-08-01