Chinese technology giant Tencent has launched its latest artificial intelligence model, Hy4, with built-in safeguards designed to reduce the risk of misuse in cyberattacks and harmful biological research.
The company says Hy4 was trained using cybersecurity data focused on defensive tasks, such as identifying software vulnerabilities and helping developers fix them. Training material related to offensive activities, including developing exploits and breaking into computer systems, was excluded.
Focus on Safer AI Development
Tencent said its internal security assessments found no significant concerns about Hy4's ability to support cyberattacks in the areas tested. The model also offers visual understanding, coding, chart analysis and computer-use capabilities.
The company plans to release the model's weights and technical documentation, supporting its open-source development approach.
Growing Focus on AI Safety
The launch comes amid increasing concerns about AI-powered cyber threats. Tencent's move highlights the growing efforts among technology companies to develop advanced AI systems while introducing safeguards against potential misuse.
Published: Saturday, 7 November 2026 | 12:19 PM IST