Headline: Anthropic Co-Founder Says Mandatory AI ‘Kill Switches’ May Be Needed

Date: 14 September 2026

Tags: Artificial Intelligence, AI Safety, Technology, Regulation

Summary:

• Anthropic co-founder Jack Clark said AI companies may eventually need mandatory mechanisms capable of completely shutting down dangerous systems.

• Clark said most leading AI laboratories, including Anthropic, already have methods allowing them to shut down their models.

• He suggested independent third parties could potentially verify whether companies have effective emergency shutdown mechanisms.

• Clark said allowing artificial intelligence to remain a “totally unregulated industry” would create unacceptable risks.

• Anthropic chief Dario Amodei has called for AI development to slow and face greater monitoring.

• Anthropic researcher Evan Hubinger recently estimated a greater than 10% chance AI could cause human extinction within a decade.

• Nobel Prize-winning computer scientist Geoffrey Hinton said a 10% extinction risk estimate was “not unreasonable”.

• US lawmakers have proposed legislation requiring companies to maintain shutdown mechanisms for potentially dangerous AI systems.

• President Donald Trump opposes attempts to slow AI development and has dismissed claims that AI could destroy humanity.

• The UK government recently rejected a national AI kill switch, arguing dangerous models could still operate or be developed overseas.

Why this matters: Senior AI industry figures are increasingly calling for regulation and safeguards as artificial intelligence systems become more capable and autonomous.

What’s next: Policymakers face decisions over whether shutdown mechanisms should become mandatory and independently verified across the AI industry.