Z.ai and Concordia AI propose six stages for managing open-weight AI risk

Chinese AI company Z.ai and Beijing safety consultancy Concordia AI have published a framework for managing the risks of open-weight AI models, built around six stages from risk identification to governance. The authors cite three core problems: releases are hard to reverse, fine-tuning can strip safeguards, and developers lose visibility once weights are public.

• The framework sorts models into green, yellow and red zones: full release, restricted or staged release, or suspension
• Risk thresholds are defined along four dimensions: where the model runs, who might misuse it, what it enables and how well society absorbs harm
• Z.ai's GLM-5.3 is the example: vetted partners tested it first, full weights released only after safety evaluations
• The report is a companion to the Frontier AI Risk Management Framework 2.0 from Shanghai AI Lab and Concordia AI, published in July