Millennials are uniquely unconcerned abut AI

I was reading an article in the Economist recently about, among other things, attitudes towards AI. Every article in the Economist these days is about attitudes towards AI in one way or another, and I usually let it wash over me, but this one included a little chart that I haven't been able to stop thinking about.

It was a comparison by demographic group of how many people think the risks of AI are very serious — and specifically how those numbers have changed between the end of August, before all the most recent revelations about Claude and ChatGPT agents hacking things without their creators realising, and three weeks later, once all that news had sunk in.

Almost across the board, no matter how you cut and sliced things, people were (understandably) more worried the second time around. There were variations, to be sure: women had grown more concerned than men, Democrats more concerned than Republicans, and so forth, all the patterns falling along the familiar ideological fault lines. But in general, opinion over those three weeks definitely shifted towards more concerned, across every sub-group that the chart split out.

Every sub-group, that is, except Millennials. While people under 30 (“Gen Z,” for ease) and people over 45 (“Gen X,” likewise) jumped ten percentage points in terms of how many people said the risks of AI were very serious, the chunk of people in the middle, aged 30-44, were actually a few percentage points less likely to say so the second time round — even after three weeks of increasingly grimmer revelations about what AI agents had been getting up to.

Naturally I find this interesting because I’m a Millennial myself (barely), and I certainly don’t feel less concerned about AI than I did three weeks ago — though perhaps, admittedly, I don’t feel much more concerned either. But more to the point I'm curious why my generation is apparently so much more blasé about it than the ones on either side of us (or, apparently, anyone else). Whence such indifference?

One must be cautious, of course, about speculating too glibly . If I learned anything from my sociology degree — and in fact I learned many — it’s that it’s very easy to come up with convincing ex post facto explanations for survey findings, even if those explanations have no factual basis. And certainly there are lots of pat “truths” about Millennials that you could reach for to explain this particular survey finding. For example, that our formative years were bookended by 9/11 and COVID, and in the middle contained the war on terror, the financial crisis, and Brexit, to name a few — and that we’re therefore so inured to major social upheaval and existential threats that new ones no longer really move the needle.

Or, perhaps, we could theorise that Millennials have already priced in AI fear: that we can’t become more concerned about AI because the needle there is already at ten out of ten. After all, unlike Gen Z, who are just entering the job market at the bottom of the ladder, and unlike Gen X, who are already comfortably at the top of the ladder or even eyeing up retirement, we Millennials in middle age and middle management are uniquely vulnerable at work, to a technology trained to replace our expertise using digital content we probably created ourselves.

Or maybe there’s a point about familiarity and witnessing change, too. Much is made of the fact that Millennials are the only generation who experienced the rise of the internet and the smartphone as part of their childhood, and that growing up with that change is what made us the original “digital natives.” (I put scare quotes around that because the Millennial version of being a digital native means knowing how to use Excel and set up a printer, neither of which are particularly helpful digital skills these days.) That might explain why Gen Z are so much more concerned about AI — because like Millennials, they’re growing up with AI and so are the first “AI natives,” fully cognisant of its true dangers — while Millennials are blithely dismissing it because we’re over-confident in our understanding of technology.

All of these possibilities sound more or less convincing, and doubtless there are dozens of others I could come up with as well. But like I said, the fact they sound convincing doesn’t mean they’re correct, and if I really want to explain that Economist chart I'd need a hypothesis I can test with data. So, let’s look at those explanations in turn.

First, the idea that Millennials are so inured to existential dread that nothing moves the needle. Well, if that were true, we’d also expect to see polls about other existential threats showing the same Millennial nonchalance. But are Millennials similarly unconcerned about the war in Iran? About Donald Trump’s attempts to completely unchain executive power? No and no. In a recent national poll by Emerson College (my alma mater but a respectable pollster nonetheless), Millennials were about level with both older and younger respondents in terms of opposition to the Iran war. They were also the least likely to have neutral or no opinion about Donald Trump’s presidency, and the most likely to disapprove of it. On the most analogous question (“How likely do you think it is that a world war breaks out in the next four years?”), the largest portion of Millennials answered “very likely”, and that proportion was second only to Gen Z among other generations. Clearly, our needles can still be moved, and indeed are still hovering at the high end of the scale on many things.

What about the second option then, the idea that Millennials aren’t becoming more concerned about AI because our level of AI concern is already as high as it could possibly be? In that case we’d expect to find polls about AI showing extremely high levels of AI concern among Millennials, and again, we don’t, far beyond that Economist chart. A recent Pew study, for example, showed Millennials bang in the middle between Gens Z and X on their level of AI concern, with only 34% of people aged 35-49 saying they were mostly concerned about AI. The ages don’t exactly match up with the Economist chart, but overall the Pew one tells a very similar story: that Millennial fear of AI basically hasn’t shifted from that 34% for the last 3 years, even as the generations on either side of us have stayed high or gone higher.

That leaves the third option: that Millennials are so confident in our digital abilities, we don’t believe there’s any technology we can’t ultimately control. If that were the case, we might expect to see evidence of such digital chutzpah in polls about digital fluency or comfort, with Millennials scoring themselves very high on their own digital abilities. It’s worth being especially clear about what we’re looking for here, too: not actual digital literacy, which may well be higher among Millennials but wouldn’t necessarily affect their views about AI, but rather self-reported digital confidence — particularly if this comes with evidence that Millennials consistently overestimate how good they are with technology relative to other age groups.

As it happens, Ofcom recently conducted a study looking at exactly this issue: confidence versus ability at completing digital tasks, by age group. But here, too, the evidence is equivocal at best. Millennials did indeed express very high levels of confidence about their digital abilities — around 93% said they were confident internet users, and around 80% said they were confident at evaluating the truthfulness of online information — but this confidence was not any higher than Gen Z. More to the point, while Millennials in the same study did better at identifying scam emails than Gen Z, they were consistently worse at it than Gen X and even older generations. That does point to digital overconfidence among Millennials, for sure, but not enough to explain their lack of concern about AI relative to Gen Z — because Gen Z are equally overconfident, if not more so.

That’s three strikes and out for my armchair theorising; none of the easy stereotypes about Millennials really seem to explain why that Economist study found them so relatively unconcerned about AI, or why the Pew study similarly found our attitudes about AI so unchanging over the past few years. So what’s left?

Well, one thing that does jump out from that Ofcom study is that Millennials are more likely than either Gen Z or Gen X to say they’ve never heard of AI. It’s not a huge difference, to be sure — 8% of Millennials versus 5% on either side — but it’s a difference nevertheless. The other place where they pop out relative to their peers on either side is what they’re most likely to have used AI for, if they have heard of it: marginally less for research, and marginally more out of idle curiosity. These are still really small differences, perhaps not even enough to be statistically significant, but they’re the only places where Millennials stick out from what’s generally a pretty clear directional trend going from young to old.

Again, it’s dangerous to ex post facto try and explain that difference with explanations that “sound right”, and — another chestnut from my sociology degree — correlation doesn’t prove causality, either. But at the very least all these surveys taken together start to suggest a little constellation of factors shared by that core Millennial age group in their mid-to-late-thirties and early forties: that they’re not that interested in AI, don’t use it as much for work, and don’t take it very seriously. We're not concerned because we're not really paying attention.

Why that might be, it’s still hard to say. But maybe just recognising the trend is enough. AI may well turn out to be yet another bubble that eventually bursts, and perhaps I’ll yet look silly for suggesting otherwise. For now, though, as much as Millennials might enjoy acting smug and making fun of anyone who takes anything too earnestly, it seems like we should all start taking AI more seriously. I, for one, would rather not be part of the only generation that was fiddling while Claude burned Rome to the ground around us.