China’s ByteDance makes major AI safety push with global hiring drive

China’s tech giant ByteDance’s AI research unit Seed has reportedly formed an Alignment and Preparedness team and has commenced a global initiative to recruit 10 researchers in a major move to make a strategic investment in AI safety and recognize its increasingly capable models.

The recruitment surge coincides with mounting oversight of AI-agent safety grows in China and the US, particularly as developers grant them internet-access autonomy.

ByteDance’s recent announcement is a highly unusual public revelation from a dominant Chinese AI developer of a dedicated effort focused on catastrophic risk mitigation for emerging risks.

On Tuesday, state-backed newspaper The Paper reported that Seed was the first Chinese frontier AI lab to establish a safety and alignment division. As fundamental design requirements grow, safety and alignment are no longer simple prerequisites but core capabilities that can be deployed at scale.

Seed further asserted that it would equip the new team with computing capacity and research resources comparable to those available to frontier-model teams.

The positions are primarily based in Beijing, Shanghai and San Jose, California with its comprehensive AI safety auditing across model development stages, automated testing for AI security risks and preparations for future risks. In addition, Seed said it would also pursue independent validation through academic and institutional partnerships, applying continuous safety alignment during model training, evaluation and deployment.