Anthropic co-founder calls for compulsory AI kill switch, Trump dismisses fears as hoax
Anthropic's Jack Clark tells the BBC AI firms may need a compulsory kill switch, while President Trump rejects calls to slow development as a hoax.
The Full Story
A plain summary built from the channels that reported this story.
Anthropic co-founder Jack Clark has told the BBC that AI firms may need a compulsory kill switch to prevent the technology spiralling out of control. In an exclusive interview, Clark warned that AI is getting smarter and more capable every few months, and that the risks are no longer academic. He highlighted scenarios such as swarms of AI taking over the internet, closing down power stations, or wiping out financial institutions. Clark argued that the chance of dealing safely with a system a hundred times smarter than the smartest person is low, and called for national level public conversations about the technology.
The interview comes as President Trump angrily rejected calls to slow the development of AI, calling fears it poses a threat to humanity a hoax. The President's dismissal of safety concerns drew attention to a growing split within the industry, where figures such as Elon Musk and OpenAI have joined calls for some sort of slowdown. The tensions caused tech stocks to fall around the world.
Clark acknowledged that his company has systems in place to pull the plug, but said society should consider mandating kill switches and verifying them by third parties. He said it is not too late to act, but that the world is close enough to the danger zone to need serious conversations.
The debate is set against a backdrop of an intense race between the United States and China to dominate AI. China has pushed back against claims that its development poses international security risks, and President Xi Jinping has called for an international framework and more sharing of technology. Xi is due to meet President Trump in Washington later this month, and the two leaders will face pressure to agree on some form of global governance.
In the meantime, experts argue that human control is the central issue. Lisa Martin, an AI specialist, said every company needs the ability to stop AI, and called for independent testing and verification of AI systems. She noted that AI is moving from answering questions to taking actions, giving it authority over data and networks. The question, she said, is who can revoke that authority if something goes wrong.
The economic stakes are high. AI investment is boosting economies, with promises of cures for cancer and other benefits. But Jack Clark's team warned that unchecked, the economic impact could include unprecedented white collar unemployment. 'If we get this technology right, we are going to change the world in so many positive ways,' Clark said. 'And if we get it wrong, we are going to massively disrupt the economy, immiserate people and potentially cause harm.'
On screen
Stills are sampled automatically at 60-second intervals. Where shown, the still is the nearest available frame from the relevant broadcast segment and is included to show what was on screen during that part of the broadcast. A still may not correspond to the exact second of a quoted phrase.
Key Claims
Claims reported during this story's coverage, mapped by channel. Ordered by how many channels carried each claim.
| Claim | BBC News | BBC One | Sky News |
|---|---|---|---|
| Donald Trump dismissed the AI warnings as fear mongering and said a slowdown would hand the AI race to China. | |||
| Anthropic whistleblower and safety experts said there was at least a 10% chance that AI could kill all humans. | · | ||
| Jack Clark suggested that AI systems may need a compulsory kill switch to prevent the technology from spiraling out of control. | · | ||
| China dismissed US caution as malicious competition. | · | · | |
| Chinese President Xi is scheduled to visit Washington later this month. | · | · | |
| Leaders of Anthropic, xAI and OpenAI agreed that AI development should slow down and that governments should intervene. | · | · | |
| President Xi Jinping called for an international framework on AI. | · | · | |
| Shares in some AI companies fell as a result of the AI warnings. | · | · |
Channel Perspectives
What each channel focused on, with key quotes.
This broadcast leads with the exclusive BBC interview with Anthropic co-founder Jack Clark, focusing on his call for a compulsory kill switch and the risks of AI. It also includes analysis from the BBC's China correspondent and tech editor, placing the story in a global context of the US-China AI race and the economic impact. The tone is measured, presenting both the benefits and dangers of AI.
- “I think most labs have different ways of being able to pull the plug and yes, of course, we have systems in place for this.”
- “President Trump, on the other hand, did not, criticising the company, saying fears were a hoax and that the only control needed was him.”
- “Nobody I've spoken to in the industry is locking themselves in a bunker right now.”
This bulletin also leads with the Jack Clark interview, but places greater emphasis on the reaction from industry and the debate over human control. It includes a contribution from AI specialist Lisa Martin, who argues that the biggest issue is human control and calls for independent auditing of AI systems. The tone is more analytical, focusing on the practical measures needed to ensure safety.
- “What I want to know from every tech company is, what can your AI access? What is it authorized to do and not to do? Who can revoke that authority? And can you contain it or stop it if something goes wrong?”
- “don't just take the AI company's word for it, and I believe this firmly, test it, audit it independently, verify it.”
- “But it may already be too late. It's not too late and I don't believe that we're yet at this maximal kind of danger zone. But we are close enough that we need to be taking this seriously and having national level public conversations about it.”
This later BBC News bulletin repeats the same core report and analysis, with a similar selection of quotes and expert reaction. It is almost identical to the earlier BBC News broadcast, covering the same ground without significant new information. The emphasis remains on the kill switch call and the political and geopolitical tensions around AI regulation.
- “I think most labs have different ways of being able to pull the plug and yes, of course, we have systems in place for this.”
- “When you put it all together, is it really worth it? Does it cause you sleepless nights?”
Broadcast Timeline
News broadcasts tracked for this story, in time order.