The Economist.
I'm John Priddow and this is Checks and Balance from The Economist.
Each week we take one big theme shaping American politics and explore it in depth.
Today, an interview with Dean Ball on whether humans can keep control of AI.
A few weeks ago, our producer asked me a question.
If I could interview anyone for an episode of Checks and Balance right now, who would it be?
An answer came to me pretty quickly.
Like many of us, I'm fascinated, concerned and occasionally confused by AI.
To help me know what to think, I've been reading Hyperdimensional, an AI policy substack written by Dean Ball.
While he's been writing his blog, Dean has moved around the AI world a fair bit.
He's been in think tank land.
He's worked in government as a senior policy advisor
for AI at the White House and co-drafter of Donald Trump's AI policy plan.
And he recently joined OpenAI to head a team which, in his words, will shape frontier AI policy within the company.
Before all that, he was a political philosophy nerd.
There are few people who combine all these strands in one brain.
We've had Dean on the podcast before, and that was a conversation I wish had been longer.
So this one is.
It seemed like a good moment to speak with him while he's still settling
into his new job and in the aftermath of a model made by his new employer going rogue in a test.