Can you tell when content is AI-written? And does it matter?

There’s a lot of AI written content on the web now. Can you tell? I think many of us can. And does it matter to you? If it’s non-fiction and you are mainly just seeking information, do you care? I think many of us probably still do. We want to experience an authentic voice, the personality of the author.

But is this desire strong and provable enough for businesses to listen? It might be risky for you or your business to be using AI written content.

If we know an article or other text was AI generated do we trust it less? Do we feel, perhaps subconsciously, that the brand producing it has cheated us in some way? So much of the human touch has already been removed from commerce. Some shoppers miss the conversations with the shop workers that were killed off by self-service checkouts, for example. The cost saving automation looks good on spreadsheets but perhaps some elusive human values have been lost.

Yet the opportunity offered by ChatGPT and other apps is immense. Endless grammatically correct sentences can be churned out at the press of a key. In contrast, writing can be unpleasantly hard work.

It’s like a new abundant resource has been unearthed. How will businesses take legitimate and wise use of it for the benefit of their clients without losing that hard-to-prove, ineffable magic of the personal touch?

Because the value of the human voice in text is likely hard to prove, I suspect more rational concerns will win out in the short term. Companies may dispense with human-written content in many areas. I think they already have.

If I am seeking a help document I don’t care if the functional steps I’m reading were AI generated. But if I’m searching for new ideas or a point of view, surely a real individual voice is better? Or if I’m trying to learn a new skill I find it valuable to model a real person showing me.

And yet, when I read Amazon reviews I find the AI generated summaries useful. As they are an aggregate of all the human POVs they somehow still carry the ring of truth.

It seems to me that even the experts don’t know where the AI revolution is heading. Or the extent to which it will replace human thought and creativity. We look to the Industrial Revolution as an analogy. Maybe you don’t care if your bread or your clothes were made by machines. Especially if they cost less. But will you care if the blogs and books you consume are machine-written? Doesn’t content benefit from a human touch in the way material goods don’t? There is a charm to hand-made goods that carry the personality of the craftsperson that made them. We’ve mostly talked ourselves out of valuing that charm. Is this also an attitude that’s soon to be rendered out of date when it comes to reading?

Perhaps writers will find a compromise and learn to integrate their work with that of AI apps. In theory it should be possible to use such apps with a lighter touch, more like a copy editor than a ghost-writer. Or even to mimic the voice of the author. Perhaps readers-as-patrons apps like Substack, where writers can be directly paid for their work, are a sign that there is still a market for the real human voice. But its a more expensive model for the reader. A bit like how vinyl records are far more expensive now than when they were the mainstream.

Or maybe economic concerns will make human writing mostly unaffordable. And in a game-theoretic way we’ll reach an equilibrium where because many companies use AI written content all their competitors must too.

The spread of AI content feels like the moment in Fantasia when Mickey Mouse learns to conjure the brooms into carrying pails of water. Yet, ominously, in the classic Disney animation it ultimately ends in an unstoppable flood.