Imagine everyone who submitted a Thesis Statement was in a room together, trying to come up with one singular prediction that reflected everyone’s views. It’d be nearly impossible, even for some of the brightest minds in tech and business.
In my experience, AI has made it easy for one person to do the work of 10. It hasn’t made it any easier for 10 people to agree. Let alone 100, or 1,000.
Almost every AI tool today is in “single-player” mode. Whether you’re prompting one model or directing a team of agents, you’re working alone, using your own taste and judgment. As execution gets easier, the harder challenge becomes agreeing on a direction and moving there together.
That only gets harder as the tools improve. When execution is effortless, human value concentrates in taste, and taste is stubborn. The people with the clearest vision of what they want are often the least willing to compromise.
A smaller team can at least hash it out face-to-face. But when the “multi-player” scale grows—say, to the size of a town hall—then the decision making ends up belonging to the loudest voices in the room. Everyone else goes home unheard (and unhappy).
This is where AI can help: not by replacing human dialogue, but by helping us find hidden consensus and understand differing viewpoints. It can let thousands speak at once and see where they agree and where they split. Nobody waits their turn, and the loudest voice no longer sets the terms.
I’ve seen AI improve civic engagement in Kentucky and Oklahoma as communities plan their futures, and in national conversations across bitter political divides. The technology won’t settle every disagreement, but it can make each one clear enough to understand, navigate, and work through.
After automation, the bottleneck will be how well we work together. We spent a decade teaching machines to talk to us. The next job is teaching them to help us talk to each other.
