Home / Technology / Tencent Censors AI for Cybersecurity Defense
Tencent Censors AI for Cybersecurity Defense
4 Nov
Summary
- Tencent released Hy4, a new open-source AI model.
- The model was trained using limited cybersecurity data.
- Tencent filtered out exploit development content during training.
Tencent recently unveiled Hy4, its latest open-source artificial intelligence model. The company emphasized that Hy4's training process deliberately excluded extensive cybersecurity data. This selective approach was taken to prioritize defensive capabilities and minimize the model's potential to assist in cyberattacks.
During Hy4's development, Tencent filtered out training materials covering areas such as exploit creation and reverse engineering. This strategic exclusion aims to limit the model's offensive potential, a concern heightened by recent global AI-driven cyberattacks. The company stated its commitment to responsible practices in AI training and release.
Tencent's initiative mirrors similar efforts by other AI companies, including Moonshot AI, which have recently curbed their models' cyber-offensive capacities. This trend reflects a growing global awareness and concern regarding the risks posed by advanced AI technologies.
