Hello discoverers!

I was ready for bed when my partner put on The AI Doc on Netflix, which I hadn’t heard of before. Naturally, she was asleep within ten minutes and I ended up watching the whole thing. Not completely unexpected, it turned into an irritating late-night rollercoaster ride through humanity’s potential futures.

Directed by Daniel Roher (of Navalny fame), the documentary explores the existential risks and potential promises of AI through a colourful collage of interviews, archival news footage, whimsical sketches and animations. Visually, it’s a joy to watch. He talks to many of the big names in AI, vocal sceptics but also the CEOs themselves, while making himself – a soon-to-be father – the film’s central character.

The first third is pure doomer fuel, talking about catastrophic misuse and the end of our species. He then overcorrects with a techno-optimist counterweight – a parade of accelerationists describing AI as the beginning of a new age of abundance and human potential. That part made me so furious that I had to get up and start stacking the dishwasher to avoid yelling at the TV.

It’s quite a tonal journey: thanks to AI, your kid might not make it to high school, but also, congratulations, he’ll thrive without ever needing a job. Roher eventually tries to land somewhere between these two extremes, and – surprise! – it’s not a place that gives us much closure.

The documentary aims for broad appeal rather than going particularly deep (for depth listen to Ezra Klein’s monologue below), and it leaves plenty of important territory underexplored. Many of the interviewees also feel too lightly challenged. It’s a very US-centric, very male group of ‘experts’.

But that’s partly beside the point. Roher isn’t a journalist. The real subject is a human quest: how frightened should we be about the future we’re building? Is it worth bringing children into it? The underlying theme is one of dread, and I’d say most of us carry some version of it.

I won’t spoil too much about how he attempts to resolve the tension, but there was a line in the closing that I found hard to disagree with: AI is too important a technology to simply be cynical about.

Cynicism has sadly become my default response to things coming out of Silicon Valley. It’s tempting to leave the whole thing to burn itself out while we make snarky remarks and carry on with our lives.

But wherever you are on the doomer–cynic spectrum, disengagement is too big a gamble. We don’t have to believe the hype, trust the companies behind it or like where things are heading. But we do need to understand the technology well enough to push back, and to make sure the people building the AI future aren’t the only ones deciding what it looks like.

The film only touches on this briefly, but nuclear technology could offer a useful precedent: we didn’t stop developing it, but we eventually agreed that something this powerful shouldn’t be left entirely to the people in the labs or military. Public pressure helped push governments towards testing bans, tighter safety rules and independent oversight. AI needs the same.

For all its shortcomings, The AI Doc does what a documentary should: it makes you feel more strongly about an issue and leaves you with something to argue about. Which is good, because, once again, we need enough people to pay attention and make it politically inconvenient for governments and companies to do whatever they like.

And now to this week’s discoveries. – Kai