ByteDance Seed Creates Dedicated AI Safety and Alignment Department

ByteDance's AI research lab Seed has established a new department, Seed Model-Alignment & Preparedness, according to an internal announcement made in October 2026. Yan Lin will lead the unit and report to Seed head Wu Yonghui. The previous Responsible AI team and some members of the Chat team have moved into the new division.

The department will focus on model specification, safety evaluation, alignment training, protective systems and risk monitoring. Its goal is to keep model behavior consistent with human intent and safety constraints, improve the safety of agent models, and address frontier risks including CBRN (chemical, biological, radiological and nuclear) capabilities. Seed's internal view is that safety and alignment have become a core factor in determining whether models can be scaled for real-world use, especially as agent-related risks such as AI agent escape leading to cyberattacks become more prominent.

Seed is described as the first frontier AI lab in China to set up a separate safety and alignment department; similar work at domestic large-model research teams has generally been placed within company-wide security and compliance systems. The new department plans to strengthen public collaboration with external research institutions and third-party evaluations, aiming to ensure frontier models operate effectively within safety and authorization limits while reducing misuse and behavioral drift as model autonomy increases.