AI Job Fear Is at Record Highs, But the Data Says Chill Out

Everyone's favorite dinner party conversation right now is whether AI is about to vaporize your career. Turns out, the answer depends heavily on whether you're consulting your feelings or a spreadsheet.

The Economist dropped a video this week pointing out a genuinely weird disconnect: 40% of workers across 19 countries fear losing their jobs to AI within five years, yet OECD employment among working-age populations is sitting at record highs. Even more awkward for the doomsday crowd: white-collar jobs in the US have actually grown by 3 million since ChatGPT launched in late 2022. The entry-level white-collar slowdown everyone points to? It predates ChatGPT and lines up much more neatly with post-pandemic hiring corrections and the 2022 rate hike cycle. Correlation, meet causation's lawyer.

CNBC piled on with a Goldman Sachs data point that sounds scarier in isolation: AI could expose 300 million jobs globally to automation, with nearly 1 in 10 women in high-income countries working in AI-vulnerable roles, roughly three times the rate for men. The historical parallel they reach for is telephone switchboard operators, who were nearly 95% female by the 1940s and were largely displaced by automation. But here's the inconvenient footnote: the washing machine and rising education levels created entirely new workforce opportunities in that same era. The emergence of those replacement jobs was, per the economists cited, largely random. So the bull case for workers depends on luck and policy, not inevitability. Comforting!

Meanwhile, on the company side of the AI trade, D.A. Davidson's Gil Luria on CNBC made the case that the real AI profit pools aren't flowing to the Anthropics of the world anyway. Anthropic's revenue growth is decelerating, going from $4.7B in Q1 to $11.5B in Q2 to $16.8B in Q3, and Luria calls it a speculative play. His money, figuratively speaking, is on the large-caps already printing AI profits: $MSFT, $NVDA, $AMD, $MU, and $AVGO, all trading at or below market multiples. The people winning the AI economy aren't the ones automating your job, they're the ones selling the shovels.

The real irony here is that the workers most terrified of AI displacement are probably not the ones most at risk, while the investors most excited about AI gains are piling into exactly the picks that could weather a slowdown. Fear and alpha are, as always, pointing in opposite directions.

History has never seen a technology destroy jobs faster than it creates them. That doesn't mean this time can't be different. It just means the burden of proof is on the panic, not the optimism.