AI Image
Tencent AI chief warns models will soon pose biology risks, SCMP reports
The Next Web • • business
Key Points:
- Tencent’s chief AI scientist, Yao Shunyu, warned that AI models will soon have the capability to cause harm in biology, similar to existing threats in cybersecurity.
- Yao emphasized the importance of safety measures for Chinese companies releasing open AI models, noting that biological risks may be more difficult to manage than cyber risks.
- Tencent recently launched its flagship AI model, Hy4, which excludes exploit-related data and focuses on defensive security, with strong basic biology knowledge but limited experimental design capabilities.
- Tencent plans to release Hy4’s model weights and a safety testing report soon; Hy4 is currently ranked as the second-best Chinese AI model by benchmarking firm Artificial Analysis.
- Yao suggested a balanced approach combining open and closed AI models, recognizing the security benefits of openness while acknowledging the arguments for keeping top models closed, as seen with US firms like Anthropic.
