A month ago, Washington and Beijing agreed to set up a channel to alert each other in the event of a dangerous AI incident. It echoes the Cold War when Washington and Moscow, in 1963, set up a hotline to prevent accidental nuclear war after the Cuban Missile Crisis.
In many ways today’s AI race carries the same volatile mix that defined the nuclear race: a transformative technology racing forward, nagging mistrust, a deep fear of falling behind that keeps anyone from slowing down, and outsized, existential risks if something goes wrong.
The nuclear race has lessons in how we navigate the AI race. Establishing a hotline is one example. And it’s the sort of step that Tristan Harris has urged. The film featuring him, The AI Doc, is now on Netflix, and this month, ahead of the screening of the film in Reykjavik, Iceland — where Reagan and Gorbachev met in 1986 to negotiate nuclear arms reductions — he called for another moment like that: “We have to make that happen again right now.”
I’m recovering this week from surgery to repair a broken collarbone that refuses to heal, so I’m sharing this conversation I had with Tristan at Faena Rose in June. Of all the problems we’re solving for today, none is bigger than making AI safe. In this conversation, we talk about how to do that.
We’ve spent a good bit of time exploring the solutions to AI safety. For a deeper read, check out my previous series on AI safety and, in particular, my piece AI: The New Nuclear Moment. Also, When Society Becomes the Experiment and The Week AI Safety Got Real.
Solving For is for people who want to dig into hard problems and find ways to solve them — unpacking the stakes, examining the forces, and surfacing paths forward. We’ve published 11 series to date on topics ranging from Rare Earth Elements and Space to the American Dream. Each unfolds in three parts. You can read or listen — narrated by me — at solvingfor.io. Learn more.
CHECK OUT PREVIOUS SERIES:
