Long before any of us had heard of ChatGPT, researchers at the Berlin Institute of Technology put a group of engineering students in charge of the life support on a spaceship (a simulated one, thankfully). Each student had an automated assistant to flag any faults, and each person was warned that the assistant could miss things. Half of them were also shown the assistant missing a couple of faults during their training. When a fault later slipped past it in the test, they were the ones who mostly caught it in time, while most of the students who had only been warned missed it.
I'll keep coming back to those students, because every one of us now sits beside an automated assistant of our own, and most of us are behaving like the half who were only warned.
I can say that with some confidence because people have counted. When King's College London surveyed UK workers who use AI, fewer than half said they usually check what it gives them. In an experiment at Wharton, people went along with an AI's wrong answers about four times out of five. We've always been a soft touch for a confident bullsh*tter, and this one also happens to tell us that our ideas are great.
What worries me is what that over-reliance does to us over time, because thinking is a skill like any other, and a skill you stop using starts to fade. In one Polish study, doctors with an average of 28 years' experience were given an AI to help them spot growths during colonoscopies. Within three months they were spotting them less often when they worked without it. (It's one study and it can't prove the cause, but it's what you'd expect.)
The big consultancies have clocked this, and since they sell expertise and thinking for a living, it's a serious problem for them. This month PwC's US firm said it will spend more than a hundred million dollars teaching 85,000 of its people to think critically. EY's US firm has put a similar sum into bonuses for skills like judgement. When every firm has the same AI, the quality of your people's thinking is what's left to tell you apart. That makes critical thinking the USP of the future, and something you either protect or watch fade.
So if you're a CLO, the question on your desk is how you help thousands of busy people protect and build their critical thinking when the shortcut is sitting right there on the screen. I'll save you some time by saying that another course won't do it. I set out why in Death by a thousand Tuesdays, which was about ethics and turns out to apply just as well to critical thinking.
I say that because your people have already been warned, just as every one of those students was. They know perfectly well that they should check, and when the deadline is close they press send anyway, because humans are gonna human, as my colleague Paul Mallaghan likes to say. That makes this a behaviour problem, and in my experience behaviour only shifts when people feel why something matters to them.
The students who caught the fault had lived through it once already, and that turned out to be worth more than any warning. Nobody wants their people learning that lesson on real client work, which is exactly where story earns its keep. Not story as decoration. Not a film on the intranet. Story as the deliberate use of narrative and experience to help build judgement, confidence and capability when it matters most.
A good use of story puts you inside the moment before it happens to you. You watch someone a lot like you take the machine's word for it and get it badly wrong in front of a client, and your stomach drops along with theirs. Then the next time you're the one with a deadline and a plausible answer on the screen, something familiar tugs at you. It's the feeling that you've been here before and it didn't end well, and that the better way through is also the one that protects you.
That's the work we do at Tilt, and the clearest example I can give you has nothing to do with offices. When Smashed, the underage drinking programme run by Collingwood Learning and backed by Diageo, needed to go online, we built it as a interactive film in which teenagers make the choices themselves and see what follows. Of the young people surveyed afterwards, 92% said they were less likely to drink underage.
I should say that I love this technology and my team uses it every day, so please keep using it for the right reasons while you look after the one thing it can't replace. Your thinking is your product, after all, and it's the same product the big consultancies are now paying very good money to protect as the reality of the AI race kicks in.
If you're working out how to do that for your own people, I'd love to compare notes.
AI Disclaimer
Research Links - sourced by AI, verified by hoooman.
Insights about consultancies and their links - sourced manually.
Article. Structure, drafted & written by Jonathan Malyon.
Grammar checked by Claude.
