This week Barry Phillips looks at research suggesting AI is helping individuals but damaging teams. 

Transcript: 

Hello Humans! And welcome to the weekly podcast that aims to address a key development in AI relevant to HR in around five minutes.

My name is Barry Phillips

Picture this. A Formula 1 car rolls into its pit box and stops dead. In an instant, twenty people swarm around it. Wheel guns scream. Four tyres off, four tyres on. A couple of seconds later, it's gone.

Watch closely, and notice what nobody does. Nobody checks anyone else's work. The front-left crew never glance at the rear-right. They don't need to. Everyone knows exactly who's doing what.

Every team you support has its own pit crew. So, today's question is this: what happens when a brand-new member joins the crew? One who never sleeps, never says no, and seems to have an answer for everything.

Yes, we're talking about AI. Not what it does for productivity, but what it does to the invisible wiring between your people.

Back in the 1980s, the psychologist Daniel Wegner gave a name to something we all do without thinking: transactive memory. The idea is simple. Groups share out the job of remembering. Nobody needs to know everything. They just need to know who to ask.

Good teams carry a mental directory. It's exactly what lets a pit crew skip the double-checking.

So, what is AI doing to that directory? New research from Kate Niederhoffer and colleagues, reported in Harvard Business Review this week, gives us a fascinating answer. They surveyed over a thousand full-time desk workers in the US, all on project teams, across more than twenty industries.

The good news first. The more people used AI, the more they saw their teammates as having distinct, specialist knowledge. Tick.

But here's the catch. When people turned to AI instead of a colleague, things started to go wrong. In teams where that had become the habit, trust and coordination each fell by around 22 percent.

The researchers call that swap 'displacement'. And they've coined a word for what follows: 'relationslipping'. The slow loosening of working relationships, one skipped conversation at a time.

Relationslipping. Not a word I expected to say on an HR podcast. But here we are.

So why does this happen? I'd describe it as two slow leaks.

Leak one: Every question you take to a chatbot is a conversation you didn't have with a colleague. Do that often enough, and you stop learning who on your team actually knows what.

Leak two: the work goes foggy. When colleagues lean heavily on AI, it gets harder to tell whose thinking you're looking at. Real expertise, or a confident copy and paste? When you can't tell, you start double-checking.

Go back to the pit lane for a second. Imagine every mechanic stopping to inspect the other three wheels. Two seconds becomes twenty. And the race is gone.

The hopeful part? The researchers don't blame the technology. They blame the choice to swap people for it. And choices can change.

Let's finish where we started. The fastest pit crews in the world aren't fast because they have better wheel guns. They're fast because every single person knows exactly who's doing what and trusts them to do it.

AI can make each of your people sharper. Brilliant. Your job is to make sure it doesn't make the team blurrier.

So, here's my challenge for you this week. Next time you, or someone on your team, reaches for AI, just pause. Ask yourself: is there a person who'd know this? And would asking them to make us stronger?

That's it for this week. As always thank you for listening. See you next week.

Until then, look after your pit crew.