This Week's Term: Centaur Teams - human-AI partnerships that combine human judgment with AI computational power, named after the mythological creature that was half-human, half-horse. The term originated from Garry Kasparov's "Advanced Chess" experiments following his famous 1997 defeat to IBM's Deep Blue.
In Issue #12, we explored Kasparov's journey from AI adversary to collaboration advocate. But there's a deeper story worth telling - one that gave us the term "centaur" and a powerful lesson about human-AI teaming.
After losing to Deep Blue, Kasparov didn't retreat. In 1998, he invented a new form of chess where humans and computers work together instead of against each other. He called it "Advanced Chess," and the human-AI partnerships became known as "centaur teams" - borrowing from the mythological creature that combined human intelligence with the power of a horse.
Within a few years, centaur teams of amateur players with midrange computers were beating both human grandmasters and top chess supercomputers. Neither humans alone nor machines alone could compete with the combination.
The most striking result came from a 2005 freestyle tournament. The winners weren't grandmasters with powerful computers. They were two amateur chess players working with three ordinary PCs running different chess programs. When the programs disagreed on moves, the humans "coached" them to investigate further. Their process for collaboration mattered more than raw capability on either side.
Kasparov's conclusion has become a touchstone for human-AI collaboration: "A weak human player plus a machine plus a better process was superior to a strong computer alone and, more remarkably, superior to a strong human player plus machine and an inferior process."
The centaur model suggests that competitive advantage in the AI era won't come from having the most powerful AI or the most talented humans. It will come from designing better processes for human-AI collaboration - exactly what we explored in the Business Value section of this issue.
There's also a counterintuitive insight buried in the chess research: experts can actually be at a disadvantage. Grandmasters sometimes "fall prey to the very human belief that as experts in their fields, they know better than the machine." The amateurs who won in 2005 had no such ego. They were better at knowing when to trust the AI and when to apply human judgment.
As you build your own centaur teams, remember that the goal isn't human OR machine excellence. It's designing the collaboration between them. The process is the product.