There’s a common assumption that AI will democratise knowledge and skills by putting everything within everyone’s grasp. Think of all the times you’ve heard of someone expressing delight at being able to use ChatGPT to write a CV, a poem or software code, or to create an image or a video.
But it seems to me that this democratisation could be a dangerous thing. Let me explain.
There’s recent research that suggests AI adds more value to the work of experts than to novices. In Artificial Intelligence, Scientific Discovery, and Product Innovation, Aiden Toner-Rodgers of MIT reports that experienced scientists use their knowledge to assess and prioritise AI research suggestions, while novices waste time testing false positives. He also notes that experienced scientists know their expertise adds value to their use of AI (and that novices say their knowledge was of little use). He concludes that his results show that “only scientists with sufficient expertise can harness the power of AI.”
This chimes with my own experience as a writer. I’ve spent my career to date doing things – by necessity – the long way.
It gives me a huge advantage when it comes to assessing the output of an AI. I can quickly spot duplicated ideas, circular thinking and irrelevancies. I know that AI is useful as an early jumping-off point and it’s invaluable for checking and refining text at the end of the process, but it’s very much my assistant and not vice versa.
But here’s the thing.
If I was starting out in my career today, without my years of experience to draw on, how would I be able to judge the quality of an AI’s output?
The chances are, I’d probably assume it was OK. I’d accept what AI gave me, and move on. This is especially the case in a busy workplace where there’s always a lengthy To Do list to tackle.
I’d be more productive, for sure. But would I be more effective? Would I get better results for the company I was working for? I don’t think I would. Because when you write something yourself, without the aid of AI, you’re forced to do the hard thinking for yourself. You’re forced to confront the gaps in your knowledge and the gaps in your argument. (For more on this, see The act of writing is the act of thinking and also Writes and Write-Nots). The result is you end up with a better, more valuable, piece than the one that AI gave you in a second.
But the problem of outsourcing your thinking doesn’t stop there.
When you write things yourself, you learn from your experience and you gain expertise that’s invaluable as you climb the career ladder. If you can get an AI to give you an instant answer, how will people starting their careers today gain the expertise and experience they need to develop in their careers? And where will companies find the experienced and expert people they need at senior levels?
It seems to me that shortcuts we’re able to take now have the potential to cause serious harm further down the line. Because as the generation of people who’ve had to learn and think the long way leave the workplace, who’s going to replace them?