China Open Source Daily — 2026-10-18
Two Chinese national standards now converge on open-source AI models from opposite directions — TC260’s Guidelines for the Security of Open-Source AI Models (plan 20264767-T-469, lead drafting Zhejiang University) moved into formal drafting on 2026-09-30 while GB/T 48110-2026 on open-source model platforms takes effect 2026-12-01 — recasting open source from a licence-permission regime into an infrastructure-disclosure regime, with the AI Safety Governance Framework 3.0 having already named the notification duty and prohibited-behaviour list the draft inherits.