Tencent strips cyber-offense from Hy4 model, following Moonshot's Kimi K3.1 Pro move
pstAsiatech · x · 2026-11-16
According to Asia Tech, Tencent said earlier this month it had stripped its latest Hy4 model of cyber-hacking capabilities while retaining cyber-defence capabilities, following a similar October move by Moonshot AI with its Kimi K3.1 Pro model. These are voluntary measures by Chinese AI firms to curb open-weight model misuse. Open question: given continuing AI progress and competition among firms, whether this approach is sustainable in the long run.
More from Safety
- macOS 27.2 patches 26 CVEs including Wi-Fi flaws enabling remote system crashes — AccBalanced · 2026-11-16
- OpenAI Foundation board member: AI accidents will soon become humanity's leading cause of death — ThereWas · 2026-11-16
- Prediction: rogue AIs hijacking compute within 6 months, but 'mildly annoying' like the common cold — willdepue · 2026-11-16
- AI summary of data center permitting reveals a dizzying list of required reforms — PTrubey · 2026-11-16
- Can Chinese AI firms curb open-weight cyber risks without losing their edge? — pstAsiatech · 2026-11-16
- tszzl: Frontier AI labs already need federal sign-off before launches—an unofficial licensing regime — tszzl · 2026-11-16