China's Z.ai Delays GLM-5.3 AI Model Release by Two Weeks Amid Safety Concerns
Chinese AI laboratory Z.ai has announced a two-week postponement of its GLM-5.3 artificial intelligence model launch. The model has demonstrated strong capabilities in identifying and exploiting security vulnerabilities, drawing scrutiny similar to that faced by leading US AI firms. Z.ai claims GLM-5.3 outperforms several American models on key benchmarks and positions it as a tool for protecting open-source environments. Despite its potential defensive applications, the company has acknowledged ongoing concerns about managing the model's use responsibly after its eventual release.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in