I watched Unabomber: In His Own Words on Netflix recently.
I knew the story of Ted Kaczynski, but I had never really engaged with his 1995 manifesto, Industrial Society and Its Future.
Obviously, Kaczynski was a murderer and his actions were abhorrent. But some of his observations about our relationship with technology are uncomfortable to read 30 years later.
His argument wasn't simply that technology was bad.
He was concerned about what happens when technology moves from being something we choose to use to something we become dependent on.
New technology normally starts by making us more capable. It saves time, solves problems and improves productivity. Then businesses reorganise around it, competitors adopt it and expectations change.
Eventually, something that was optional becomes very difficult not to use.
It is hard not to see the parallel with AI.
Imagine wearing an exoskeleton for a year.
It makes you stronger, faster and more capable. You can lift twice as much, walk further and work harder.
It feels like augmentation.
But imagine that while you were wearing it, your own muscles were hardly being used.
After a year you remove the exoskeleton and discover your muscles have atrophied so badly that you can no longer stand on your own.
Were you actually becoming more capable?
I think that is one of the most important questions we need to ask about AI.
We are already handing over more of the cognitive work we used to do ourselves.
Write this.
Summarise that.
Analyse these numbers.
Research this problem.
Recommend what I should do.
None of those things is inherently bad. I use AI every day and think its potential is extraordinary.
But there is a big difference between using AI to augment our capability and gradually surrendering capability to AI.
That is why I think the old idea of "teaching someone to fish" is important.
Good AI adoption shouldn't just give people better answers.
It should help them become better at asking questions, framing problems, challenging outputs and understanding when the technology is wrong.
People need to understand AI well enough to use it wisely, rather than blindly.
Otherwise we risk creating organisations that appear increasingly capable while the humans inside them progressively lose the ability to operate without the technology.
This also connects with something I wrote about recently: the Adaptive Organisation
My argument there was that competitive advantage in the AI era won't come simply from adopting AI faster.
It will come from an organisation's ability to continually adapt how people, technology and work fit together.
Sometimes that will mean automation.
Sometimes augmentation.
And sometimes it will mean deliberately keeping a human capability because judgement, experience or understanding matters.
For me, that also changes the role of people helping businesses adopt AI.
Our job shouldn't be to become the person who keeps catching the fish.
It should be to transfer capability into the organisation.
Teach people how the technology works.
Teach them where it fails.
Teach them how to challenge it.
And then help them redesign work around what humans and machines are each best suited to do.
AI literacy isn't just about productivity.
Increasingly, I think it is about preserving human and organisational capability.
Because if the exoskeleton ever comes off, we should still be able to stand on our own.