United States
United Kingdom
Canada
Australia
New Zealand
IrelandSteven Bartlett's AI panel divides over the point of no return
The Diary Of A CEO Clips excerpt pits the benefits of continued development against the warning that waiting for proof of danger could mean waiting too long.
The Diary Of A CEO Clips
Story of Why this story? 965,239 views so far, and here is what is inside it.
If YOU Think AI Is Safe, You NEED To See This
Things we can learn from the video
Press to jump
- 1:06Bartlett asks whether a small chance of human extinction could ever justify pressing an AI-development button.
- 1:42McAfee argues that halting AI over speculative dangers would sacrifice tangible benefits already available.
- 3:12A proposed alternative is narrow AI for specific problems, rather than general superintelligence.
- 5:41McAfee offers a hypothetical harmful takeover of Waymo vehicles as evidence that would change his assessment.
- 10:40A panelist argues that superintelligence could pass a point beyond which trial-and-error corrections are impossible.
- 15:26The closing argument links software errors to unreliable control and calls for permanently banning general superintelligence.
Driverless cars become both the promise and the warning in Steven Bartlett's AI roundtable. Andrew McAfee points to automated driving's potential to prevent road deaths. Asked what would persuade him that AI posed a severe danger, he offers a hypothetical example: an uncontrolled takeover of Waymo vehicles that harms people.
That difference, between benefits available today and catastrophe that has not happened, drives the discussion in "If YOU Think AI Is Safe, You NEED To See This". The excerpt on The Diary Of A CEO Clips has drawn nearly a million views since publication.
Bartlett makes the dilemma personal with a button-pressing thought experiment: would you accept a small chance of wiping out humanity? McAfee argues that stopping development on the basis of a speculative chain of events would itself be immoral, because it would deny people AI's benefits.
The opposing argument is not simply that AI might go wrong. It is that the usual way of learning from technology's failures may stop working once a system becomes powerful enough to escape human control. Evidence sufficient to convince everyone could arrive after the opportunity to intervene has gone.
One proposed alternative is to develop narrow systems that solve specific problems rather than pursue general superintelligence. Later, the discussion turns to the commercial incentives of AI laboratories and whether their structure needs changing. By the closing exchange, the strongest prescription is a permanent ban on general superintelligence, grounded in the argument that complex, error-prone software cannot be controlled indefinitely.
Published on Trari
Where next?
What viewers say
I appreciate this format- a debate so we can see the fabric of the beliefs held by people who understand far more than I do about the topic.
@confidencecafeSee it on YouTubeThis is like watching The Fellowship of the Ring argue
@LogenJohnsonSee it on YouTube
Public comments from the video's page on YouTube
Is this your channel?For the channel's owners and team: a ready picture and text for your audience
Share the story with your audience. Everything is ready:
For an Instagram story or a WhatsApp status
For a community post or your channel's description
- Channel
- The Diary Of A CEO Clips
- Subscribers
- 2,940,000 ▲ 20,000 this week
- Channel views, last 7 days
- 10,554,060
- Channel views, pace per day
- 1,389,977 (on a normal day: 1,476,845)
- Video published
- November 4, 2026
- Video views so far
- 965,239 (7 days online)
- Video views, pace per day
- 76,518 (on a normal day: 213,925)
The numbers
The viewing figures give the scale of the audience for this AI discussion, rather than an unusual surge for the channel.
💡 Test your knowledge
Who hosts the AI roundtable featured in this clip?
Steven Bartlett hosts The Diary Of A CEO and puts the AI-risk questions to the panel.