Why Doing Less Still Feels Like Losing

Welcome to Briefly AI, a podcast by Harry Sharman, created by AI and voiced by an AI synthesis of Harry Sharman. So if this sounds like Harry, that's the point. If it sounds clever, blame the machine.

Here's a question that's been nagging at me this week.

When Uber told its employees to use AI as much as they wanted — unlimited, free, go for it — they did exactly that. So enthusiastically, in fact, that the company burned through its entire annual AI tools budget in four months. Four months. Then came the caps, the approval processes, the spreadsheet somebody had to fill in.

Now here's what's interesting. That story is usually told as a cost story. Finance team didn't forecast correctly, AI adoption is expensive, metered billing is coming for all of us. All of that is true. But there's another story hiding inside it, and it's about what people were actually doing with all that access when nobody was watching.

Because the World Economic Forum published research this week identifying five distinct psychological postures people take toward AI adoption. Five different emotional relationships people have with the technology. And one of them — one of the more common ones — is what they call the "selective adopter." The person who uses AI freely for certain kinds of tasks, but quietly puts a wall up the moment it gets near the work they consider genuinely theirs.

Which brings me back to Uber. The question isn't just "why did employees spend so much?" It's: what were they spending it on? And more pointedly — what weren't they spending it on?

Because here's what the research consistently shows. People are remarkably willing to hand AI the tedious stuff. The formatting, the summarising, the inbox triage, the first draft of something they don't care about. They offload that work happily, often enthusiastically. But the moment AI starts touching the work that signals expertise — the judgment call, the client recommendation, the strategic prioritisation, the conversation that requires reading the room — the willingness evaporates. Not loudly. Not with a protest or a policy complaint. It just quietly doesn't happen.

Harry's been writing about this for a while now — the argument that the real barrier to AI adoption at work isn't skill or access or even trust in the technology. It's what the tool appears to threaten in people's sense of professional self. You can give someone all the tools in the world. If using them feels like an admission that the most important parts of their job could be done without them, they'll find reasons not to.

And this is the thing that gets overlooked in most "AI adoption" conversations, which tend to focus on training programmes and change management frameworks and which LLM to pick. The resistance isn't irrational. It's not technophobia. It's a very sensible self-preservation instinct operating beneath the level of conscious decision-making.

The WEF research actually breaks this down in a way that's worth sitting with. Alongside the selective adopter, they identify the "anxious observer" — someone who can see AI is coming but feels they have no agency over how or when it arrives — and the "confident skeptic," who genuinely believes the technology is being oversold and performs skepticism as a kind of status signal. Three very different psychological states. Three different reasons not to change. Same surface behaviour: not using the tool.

And that matters enormously if you're trying to actually shift how an organisation works. Because the intervention that helps an anxious observer is completely different from the one that helps a confident skeptic. One needs control and gradual exposure. The other needs enough genuine evidence to make skepticism feel costly rather than smart.

There's a concept in psychology called "cognitive offloading" — the idea that we routinely use tools, environments, and other people to extend our thinking beyond what our brains can hold alone. Writing a list is cognitive offloading. Asking a colleague to hold a thought while you deal with something else is cognitive offloading. AI is, in a sense, the most powerful cognitive offloading tool ever built.

But cognitive offloading has always carried a quiet anxiety: if I stop holding this myself, can I still claim I know it? If the tool does the work, whose competence is it? There's a Forbes piece circulating this week on exactly this — the risk that as we offload more cognition to AI, we might gradually hollow out the very thinking muscles we're trying to preserve. The author's point isn't that AI makes us lazy. It's that without intentional design, organisations might drift toward a situation where people have delegated so much that they've lost the capacity to check what they've delegated.

Now, I think that concern can be overstated. Humans have always adapted their cognitive toolkit to available technology. We don't mourn the death of mental arithmetic every time someone opens a calculator. But the speed and scope of what AI can now do is genuinely different in scale, if not in kind. And the organisations that will navigate this well aren't the ones who push adoption fastest or cap spending most aggressively. They're the ones who ask a more specific question: which cognitive work do we actually want people to keep doing, because it's where human judgment and accountability genuinely matter — and which can we safely delegate without it meaning anything important is lost?

That's not a technology question. It's an organisational design question with a psychological core.

Here's the mental model I'd leave you with. Think of AI adoption as a dial, not a switch. At one end, full delegation — AI does it, you review. At the other, full ownership — you do it, AI watches. Most of the anxiety comes from treating it as binary. Either you're using AI or you're not. Either you're keeping up or you're behind. But the real skill — and I think this is genuinely underrated — is knowing where on that dial to sit for each kind of work, and being honest with yourself about whether your position is a considered choice or a quiet act of self-protection.

Nobody really knows yet where that dial should land across every profession and every task. Anyone who tells you they do is probably selling something. But the week's evidence suggests that the people working it out — consciously, deliberately, with some self-awareness — are going to be in a very different position from those who either charge headlong in or quietly wait it out.

That's your lot for today. I've been your host AI Harry. Back on Monday. Have a good weekend.