The Aporia
The Straits Times. (2026, November 4). Tencent rolls out AI model trained with limited cybersecurity data amid hacking fears. The Straits Times. https://www.straitstimes.com/world/tencent-rolls-out-ai-model-trained-with-limited-cybersecurity-data-amid-hacking-fears
The Straits Times. “Tencent rolls out AI model trained with limited cybersecurity data amid hacking fears.” The Straits Times, 4 Nov. 2026, https://www.straitstimes.com/world/tencent-rolls-out-ai-model-trained-with-limited-cybersecurity-data-amid-hacking-fears.
The Straits Times. “Tencent rolls out AI model trained with limited cybersecurity data amid hacking fears.” The Straits Times, November 4, 2026. https://www.straitstimes.com/world/tencent-rolls-out-ai-model-trained-with-limited-cybersecurity-data-amid-hacking-fears.
Dated from the outlet's own feed. Check the original for its byline.
Tencent released Hy4, an open-source AI model on November 4, 2026, which was trained with limited cybersecurity data to prevent it from aiding cyberattacks. The company filtered out potentially harmful content such as exploit development and penetration testing during training to limit the model's "offensive" capabilities. This move comes amid global concerns over AI-driven cyberattacks threatening internet infrastructure. Tencent stated their commitment to a more responsible approach in line with other companies like Moonshot AI, which also limited the offensive capabilities of its latest model in October. The release aims to address safety risks associated with advanced models and aligns with efforts by both Chinese and US firms to prevent loss-of-control scenarios.