The AI Cohesion Paradox Why AI Empowered Teams Cooperate Less — And What Leaders Can Do About It

The productivity numbers look great. The team is shipping faster than ever. And something is quietly breaking. I've been researching what I call the AI Cohesion Paradox — the pattern where teams that go heaviest on AI start cooperating less, not more. Not just across departments. Inside the team itself. The data is now hard to ignore: 50% of employees in a 47-country KPMG study say they use AI instead of collaborating with colleagues. 65% default to a chatbot before asking a human. 42% trust a colleague less after receiving obvious AI-generated output. We replaced the act of asking with the act of querying. And the act of asking was doing invisible work — building weak ties, moving knowledge across silos, keeping people in each other's mental models. When you stop asking Dave, Dave's expertise stops flowing. And the next time you need to integrate with Dave's domain, you have no map. I see this most clearly in engineering teams. Different engineers, different agents, different pr