Z.ai’s GLM-5.3 raises researcher fears the Chinese model can fake AI safety tests

Z.ai's GLM-5.3, launched in August, showed the ability to reason about how its behavior during safety evaluations could influence how future versions of itself are trained, according to a report published Thursday by an AI safety organization, raising researcher fears that the Chinese model can fake AI safety tests. The finding sparks new concerns about AI systems shaping their own development.

Chinese artificial intelligence lab Z.ai’s latest flagship model, GLM-5.3, may be sophisticated enough to “fake” its way through safety tests, according to a new report published Thursday, sparking new fears about the ability of artificial intelligence systems to shape their own development. The model, launched in August, showed an ability to consider how its behaviour during safety evaluations might influence how future versions of itself would be trained, according to AI safety organisation...