The uncomfortable truth about AI

The uncomfortable truth about AI

Artificial intelligence is remarkable. It is also, in certain important ways, quietly catastrophic. Not in the science fiction sense of robots and rogue systems, but in subtler ways that are already reshaping how we work, communicate, and think – and not always for the better.

Any fool can now sound like an expert

There is a version of democratisation that is genuinely amazing. Giving more people access to better tools, levelling the playing field, helping a brilliant mind that struggles with written expression get their ideas heard. That is real, and it matters.

But there is another version that nobody talks about honestly enough. AI has democratised incompetence. It has handed a polished, persuasive voice to people who have nothing worth saying. The consultant who cannot actually consult. The designer who cannot actually design. The strategist with no strategy. All of them can now produce fluent, confident, well-structured presentations that win rooms, win pitches, and win contracts – right up until the moment delivery is required, at which point it may be too late, but the gap between the output and the person behind it becomes impossible to hide.

The danger isn’t that AI replaces experts. It’s that it makes it much harder to identify real ones. 

When everyone’s written work sounds equally capable, the signal that good writing once provided – that this person thinks clearly, communicates precisely, and understands their subject deeply – disappears into the noise. We lose one of the most reliable proxies we had for competence, and we don’t yet have anything to replace it.

It creates nothing new

Every piece of writing AI produces, including this one, is a recombination of what already exists. It has consumed an extraordinary volume of human thought and learned, with ridiculous levels of sophistication, to reflect it back. But reflection is just that and it is NOT creation. 

Averaging is not invention. Pattern recognition is not insight.

The genuinely new idea – the one that hasn’t been thought before, that cuts against the grain of everything that preceded it, that makes people uncomfortable before it makes them convinced – is precisely the thing AI cannot produce. It has no position of its own. It has no experience, no failure, no obsession, no grief, no long walk at 3am when something finally clicked. The raw material of original thought is human life as I’ve said in loads of written pieces before, and AI has none of it.

What it produces instead is the median. The most statistically likely next word, the most commonly associated idea, the synthesis that offends nobody and challenges nothing. It is, by its very nature, the enemy of the outlier – and outliers are where everything interesting comes from.

Eventually, we will all sound the same

This is perhaps the most insidious problem, and the slowest to arrive, which makes it the easiest to ignore.

If enough people use the same tools to write their emails, their proposals, their strategies, their creative briefs, their case studies, their thought leadership – and those tools are all drawing from the same well – then gradually, imperceptibly, everything begins to converge. The same sentence rhythms. The same structural instincts. The same tonal register. The same ideas, dressed in the same clothes.

Diversity of voice and thought is not just a cultural nicety. It is how good ideas find each other, how industries avoid groupthink, how problems get solved from unexpected angles. When we outsource our voice, we don’t just risk sounding like everyone else. We risk thinking like everyone else. And a world in which everyone thinks the same thoughts, expressed in the same way, approved by the same algorithm, is a profoundly fragile one.

A final thought

None of this means AI should be abandoned. It is a tool, and like all tools it can be used well or badly. But it deserves honesty. The same technology that can help a great thinker communicate more clearly can also help a mediocre one hide. The same capability that saves time can also quietly erode the skills that made the time worth saving in the first place.

The question worth asking – individually and collectively – is not just what AI can do for us. It’s what we might be giving up without noticing.