What is it about?

AI is great at spotting patterns in data—like predicting what movie you’ll like or which route has less traffic. But can it actually make tough, forward-looking decisions the way humans do? This article says no. The authors explain that while AI runs on past data, human decisions often involve doing something new—something that hasn’t been tried or measured yet. Think of Airbnb: there was no data to prove it would work. But its founders imagined a new future, tested their ideas, and created new data along the way. That kind of leap—reasoning beyond the data and experimenting—is something AI still can’t do.

Featured Image

Why is it important?

There’s a lot of buzz about AI replacing human decision makers. But this article explains why that’s a big overstatement. When the future is uncertain and the data doesn’t yet exist, human judgment—not machine prediction—is what drives real breakthroughs. If we confuse prediction with decision making, we risk missing out on innovation, strategy, and the messy, creative thinking that humans are uniquely good at. In short: AI can help, but it can’t replace us where it really counts.

Perspectives

This paper was written to cut through all the AI hype and to actually highlight what AI is useful for - and where it fails.

Teppo Felin

Read the Original

This page is a summary of: Does AI Prediction Scale to Decision Making?, Communications of the ACM, March 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3722138.
You can read the full text:

Read

Contributors

The following have contributed to this page