ByteDance Seed forms AI safety and alignment team, seeks 10 researchers

Seed, the AI research arm of Chinese technology company ByteDance, has set up a new group called the Alignment & Preparedness team. In a recruitment notice posted on its official WeChat account on October 14, 2026, the unit said it is hiring 10 researchers worldwide and described the effort as a long-term investment in safety for its increasingly capable models.

The roles are located in Beijing, Shanghai and San Jose, California. Seed said the team would receive computing power and research resources on par with those given to its frontier-model teams.

What the team will work on

According to Seed, the group's agenda covers safety training throughout the stages of model development, automated vulnerability testing, tools for monitoring and interpreting model behaviour, testing environments for AI agents, and preparation for emerging risks. The unit also plans to seek independent testing and to partner with universities and research institutions, applying safety work across training, evaluation and deployment.

In its notice, Seed argued that as capabilities grow, safety and alignment have stopped being mere protections against risk and have become core capabilities that decide whether models can be deployed at scale.

A rare disclosure from a Chinese lab

Alignment refers to work aimed at making advanced AI systems behave in line with human goals and safety requirements, especially when they carry out complex, multi-step tasks. Public announcements of dedicated frontier-safety efforts by major Chinese AI developers are uncommon. On October 13, 2026, the state-backed newspaper The Paper reported that Seed was the first Chinese frontier AI lab to create a standalone safety and alignment department.

The move arrives as regulators and researchers in both China and the United States pay closer attention to the safety of AI agents, particularly as developers grant them more autonomy, including access to the internet. Washington and Beijing are competing for leadership in advanced AI while also discussing guardrails, safety risks and channels for communicating about major AI-related incidents.

The San Jose positions suggest ByteDance is looking beyond China for safety talent. How the new team's findings will shape the company's models and products has not yet been detailed.