Over the past few years, we’ve been blown away by the exponential improvements in computing in general, and in artificial intelligence in particular. From beating all opponents in the game of Go to driving autonomous cars and diagnosing cancer, it seemed that AI could outperform human players, drivers, and doctors at every turn.

Until it couldn’t. Government leaders and policy experts were relying on AI to help them navigate through one pandemic, natural disaster, and social crisis after another, to no avail. Despite employing the biggest supercomputers and most sophisticated datasets and complex models, scientists were unable to find speedy solutions to coronavirus outbreaks, climate disasters, or global economic shocks.

“These things are all complex, chaotic, connected systems,” says Ferrus Homm, a leading data scientist and cognitive systems advocate. “Whenever you put human behaviour into the mix, things become unpredictable!”

That’s part of the problem; the instantaneous feedback loops that now exist in the global digital society help us to maximize our personal utility. But when everyone tries to rationally optimize their health, wealth and individual happiness, the knock-on effects can be considerable at scale. On a macro level, all those butterfly wings are causing chaos. Never mind the irrational acts!

There’s light on the horizon. With more and more knowledge, slowly we can improve our understanding of how the world really works. And with more and more exposure to bad actors, we can improve how we operate and interact as social animals.

Smart systems and data can help, but ultimately it takes human ingenuity to solve very human problems. Don’t rely only on AI.