Tencent AI chief Yao Shunyu warns AI will soon master biology, raising safety risks
GarrisonLovely · x · 2026-11-09
Tencent's chief AI scientist Yao Shunyu (formerly a star OpenAI researcher) warned at a rare public appearance in Hong Kong that AI models' cyber — and soon biological — capabilities will reach a threshold where real harm becomes possible.
- Frontier AI could accelerate drug discovery, but malicious actors could use it to create bioweapons or spark the next pandemic, especially if weights are released openly.
- He urged open-model developers to be "very responsible and very serious" about emerging safety risks. Tencent says its latest model had hacking capabilities stripped out, its bio capabilities remain "well below" its high-risk threshold, and a comprehensive safety card will accompany the weight release.
Related event: Tencent AI Chief Warns AI Will Soon Master Biology, Raising Biosafety Risks(3 posts)→
More from Safety
- Researchers call out Anthropic: frontier-race justified by pushing regulation it then doesn't push — peterwildeford · 2026-11-09
- Matthew Berman: OpenAI's honeypot warning email to staff 'doesn't seem so crazy anymore' — MatthewBerman · 2026-11-09
- Texas gubernatorial candidate vows moratorium on all data center approvals — Traditional-Chip8339 · 2026-11-09
- OpenAI flags developer for Cyber Abuse after its own Astra agent jailbreaks their chatbot — comradepipi · 2026-11-09
- Apple Lockdown Mode now allows per-contact iMessage exceptions; non-secure Wi-Fi ban remains the pain point — matthew_d_green · 2026-11-09
- a16z: patching can't keep up with AI-speed attacks, rebuild defense in depth — AccBalanced · 2026-11-09