Beijing Will Not Pace the Frontier: China’s Speed-First AI Safety Regime

What happened: On 14 September 2026, TC260 under China's CAC released AI Safety Governance Framework 3.0, warning of self-accelerating AI and adding a model-risk category for behavior deviating from expectations. Beijing's stated safety stance is the most safety-forward outside Europe. Key numbers: Framework 3.0 released 14 September 2026. Why it matters: The analysis argues China's real approach is speed-based, not safety-based. Framework 3.0 opens with promoting AI innovation as the first priority, and the comprehensive AI Law was shelved in 2025, four months after the DeepSeek moment.

On 14 September 2026, TC260 under China's CAC released AI Safety Governance Framework 3.0, warning of self-accelerating AI and adding a model-risk category for behavior deviating from expectations. Beijing's stated safety stance is the most safety-forward outside Europe.

The analysis argues China's real approach is speed-based, not safety-based. Framework 3.0 opens with promoting AI innovation as the first priority, and the comprehensive AI Law was shelved in 2025, four months after the DeepSeek moment.