In an article for Fortune, Daron Acemoglu discusses why the productivity gains from AI are likely to be much smaller than most forecasters expect — and why the hype around AI could actually make the gains even smaller.