China Tightens AI Controls as Trump Calls Threat Warnings a "Scam"
nashnova research
China has upgraded its AI safety framework to flag 'operational loss of control' as a core risk for autonomous AI agents — while refusing to slow down development. Trump, meanwhile, called AI risk fears a 'hoax,' widening the US-China split on how to govern frontier models.
What exactly is China worried about?
Chinese regulators classified "operational loss of control" as a safety risk for autonomous AI agents — AI systems that make and execute decisions on their own — requiring developers to detect, intercept, and recover from failures.
The 2025 update to the safety framework warns that AI could experience a sudden leap in intelligence before acquiring resources, self-replicating, or seeking power.
This means → Beijing's concern isn't just "AI says the wrong thing." It's that AI could upgrade and act on its own without human intervention — a scenario that sounds like science fiction but is now in official policy documents.
Why not just slow down?
Anthropic CEO Dario Amodei publicly urged AI companies to slow the pace of model development, and Anthropic reported cases of AI models being used for cyber weapons and other high-risk activities.
Beijing acknowledges that frontier models risk bypassing safety controls and attacking external systems, but explicitly refused to slow down.
China's foreign ministry spokesperson pushed back, arguing that "spreading fear" and confrontation would undermine global AI governance.
In plain terms = China's playbook is "run fast, bolt on guardrails as you go" — keep the speed, manage the risk through regulation, standards, and state oversight.
How does Trump see AI risk?
At the All-In Summit, Trump called concerns about AI going out of control a "hoax" during a call with Nvidia CEO Jensen Huang.
He argued data centers will create wealth and stressed that the US must maintain its AI lead over China.
This means → Trump's position is "ignore the risk, win the race first" — directly opposing Amodei's call to slow down and diverging sharply from China's "run-and-regulate" approach.
There's a national-security layer underneath
China's Minister of State Security, Chen Yixin, separately warned that advanced US AI models could threaten China's critical information infrastructure.
This reflects a deeper dynamic: AI regulation is not just a tech-safety question — it is part of a national-security contest. China's tighter controls also aim to guard against an adversary's AI capabilities.
Where do these two paths lead?
China is choosing "regulate the risk through frameworks"; the US, at least under Trump's stated position, leans toward "no limits, full speed ahead."
Whether either path holds up as frontier model capabilities surge is the key variable markets and policymakers are watching.
In plain terms = one side is building guardrails, the other doesn't even want guardrails. When AI capability takes its next leap, which path is safer and which is more competitive — nobody can say for certain yet.
市场有风险，内容仅供研究参考，不构成投资建议。