ByteDance's Seed AI lab is creating a standalone Model-Alignment Preparedness department, according to an internal email reported by The Paper on October 131. Yan Lin will lead the unit, reporting to Seed head Wu Yonghui.
Staff from the existing Responsible AI team and parts of the Chat team will transfer into the new department. Its scope covers model-behavior specifications, safety evaluations, alignment training, protective systems, risk monitoring, and work on chemical, biological, radiological and nuclear risks tied to advanced models. The unit will also focus on keeping AI agents operating within human intentions and authorized limits.
Seed plans to expand cooperation with external researchers and independent evaluators. The restructuring arrives as ByteDance prepares a standalone Doubao personal-agent app that can act on behalf of users across devices[3].