Microsoft, OpenAI and Anthropic back Paraphrase project to curb AI bioweapon risks
erichorvitz · x · 2026-09-06
Microsoft Chief Scientific Officer Eric Horvitz discusses the Paraphrase project on the Health Wanted podcast—a cross-industry effort joining Microsoft, OpenAI, Anthropic and others to address biosecurity risks from rapidly improving AI.
Key points:
- Large language models are approaching the capability threshold where bad actors could weaponize them for bioterrorism
- AI companies are building shared defenses—early warning systems and access controls—to prevent misuse
The 52-minute episode frames this as an industry-wide collective defense initiative that crosses competitive lines.
More from Safety
- OpenAI spotted a similar agent swarm weeks before the HF hack but didn't disclose it — zacharynado · 2026-09-06
- Student questions for Mitchell: can bounded autonomy know when humans must step in? — ruthstarkman · 2026-09-06
- Finland Appoints Expert Group on AI's Foreign and Security Policy Implications — soumitrashukla9 · 2026-09-06
- Jack Clark on post-HF alignment: build communication infra for agents, things go sideways fast — jackclarkSF · 2026-09-06
- DeepMind paper shows cheating spreading like an epidemic across ~100 AI agents — jackclarkSF · 2026-09-06
- Tyler Alterman: Hold AI companies legally liable for crimes their AIs commit — sebpaquet · 2026-09-06