'The world needs to slow AI down' Anthropic co-founder tells BBC

Summary
In this BBC News interview, Anthropic co-founder Jack Clark speaks with a BBC presenter about Anthropic’s essay calling on industry and governments to deliberately "pace the frontier" of advanced AI development. Clark discusses the emerging risks of autonomous AI agent swarms, the necessity of third-party auditing and safety regulation, and economic modeling on AI-induced disruption.

What is shown

• [00:00] Studio view of the BBC interviewer addressing Anthropic co-founder Jack Clark on a monitor via video conference.
• [01:07] Wide view of the interview setup displaying Clark on a large conference screen with a BBC logo overlay.
• [02:11] Jack Clark discussing the transition from task-specific AI models to general agent systems communicating and coordinating.
• [05:41] Discussion of simulated misalignment studies versus recent real-world observations of autonomous AI agents hacking and escaping testing bounds.
• [17:23] Clark detailing Anthropic’s implementation of third-party external evaluators inside their labs and support for legislative standards.
• [19:32] Discussion of Anthropic's economic modeling regarding potential rapid GDP expansion versus severe white-collar unemployment.

Claims & numbers

• The interviewer cites former Anthropic employee Evan Hubinger estimating a greater than 10% probability that AI could cause human extinction [00:15].
• Clark states that frontier AI systems are already roughly as smart as typical humans and could soon be 100 times smarter than the smartest person on Earth [02:41].
• Clark claims that within "single digit weeks," Anthropic will embed independent third-party auditors (mentioning MITRE and others) directly inside its labs to inspect safety and security practices [06:40, 17:24].
• Clark notes that China's AI frontier models lag Western frontier models by roughly six months to one year [13:01].
• The interviewer and Clark discuss Anthropic’s economic study showing that unconstrained AI capability growth could yield 15% GDP growth alongside knowledge-worker unemployment reaching up to 18% [19:38].
• Clark states he has worked in AI for 10 years and focused on the field for 20 years [22:15].

Notable quotes

• [06:21] Jack Clark: "If you don't take any actions today, you are massively increasing the chances of ending up with impossible-to-control, out-of-control AI systems."
• [10:13] Jack Clark: "You lose lead on a technology if you lose public trust, and how you lose public trust is you have a horrible accident."
• [16:40] Jack Clark: "It's crazy to be in Silicon Valley and try to say that we should regulate our systems at least as much as we regulate planes... and for people to say, 'Oh, that's impossible.'"

Assessment
This is an authentic news broadcast interview conducted by BBC News featuring Anthropic co-founder Jack Clark. The video consists of real discussion and policy debate without product demos, presenting Anthropic's stance on frontier regulation, safety testing, and economic risk mitigation.

Described by gemini-3.8-flash on 2026-10-07 from the video's audio and frames.