Authenticity in the age of intelligence

Original content and its value

The books, articles, websites, conversations, research and other material created by people are what made these models possible in the first place. The demand for original human data is so significant that AI companies have reportedly gone as far as acquiring large quantities of physical books that were never digitised, scanning them page by page so their contents can become part of training datasets.

This points to an interesting problem. The amount of intelligence we can generate is rapidly increasing, but the supply of genuinely original human material is not increasing at the same rate.

In other words, as generated content becomes abundant, original thought may become increasingly scarce and therefore increasingly valuable.

This matters when we think about what we choose to put online ourselves.

There is a meaningful difference between sitting down and forming an idea from your own experiences, knowledge and perspective, and asking a model to generate that idea for you. AI can help articulate a thought, challenge it, research it and improve how it is communicated. But there still needs to be something original at the source.

For a company, that source is its people. Their experiences, opinions, stories and way of speaking are things competitors cannot simply replicate.

The value exchange of attention

There is another side to this that I think we often overlook: the value exchange between the person creating something and the person consuming it.

When I choose to read something you have written, there is an implicit exchange taking place. You spent time thinking about something, forming an opinion and communicating it. In return, I am willing to spend some of my time considering it.

Generative AI changes that equation.

Someone can now produce thousands of words in seconds and ask another person to spend ten minutes reading them. Technically, the information may be perfectly good. It may be grammatically flawless and beautifully structured. But something about the exchange feels uneven.

You invested seconds generating it. You are asking me to invest minutes consuming it.

As the internet becomes saturated with generated content, I think this imbalance will become increasingly important. People will become more selective about what deserves their attention, and evidence that genuine thought went into something may become valuable in itself.

User trust and the risk of AI slop

This leads to trust.

Have you ever visited a company’s website and immediately felt like you had seen it before?

The layout feels familiar. The colours feel familiar. The typography follows the same patterns. The copy is polished but says very little. Everything is technically correct, yet somehow the company behind it feels less distinct.

We do not necessarily need to accurately identify whether something was created by AI. We only need to perceive it as generic.

That creates a problem for businesses.

For decades, companies have spent enormous amounts of time and money developing distinctive brands. Their photography, writing, design, language and personality help us understand who they are and why we should trust them.

If those things are increasingly outsourced to the same handful of models, there is a risk that businesses begin converging towards the same aesthetic and the same voice.

The irony is that the easier it becomes to produce polished content, the less impressive polished content becomes.

What becomes valuable instead are the things that are difficult to manufacture: a genuine opinion, an unusual perspective, real experience, original photography, imperfect language, a recognisable voice and evidence that somebody actually cared enough to make something.

Authenticity becomes a differentiator.

The internet feeding itself

There is also a larger problem developing beneath all of this.

AI models depend heavily on human-created information. But as more of the internet becomes populated by AI-generated material, future models increasingly risk encountering information produced by previous models.

The snake begins eating its own tail. Essentially getting drunk on its own supply and providing hallucinated outputs as training data quality decreases.

Researchers are already interested in what happens when synthetic data repeatedly finds its way back into training datasets. If original information becomes diluted by generations of generated information, maintaining the quality and provenance of training data becomes increasingly important.

This is one reason why identifying where digital content comes from may matter more in the future.

There are already efforts to establish provenance for AI-generated and digitally modified content. Exactly how search engines, AI companies and other platforms will eventually treat this information remains uncertain.

I would therefore be careful about claiming that AI-generated websites will automatically rank lower in search results. We simply do not know.

But the broader direction presents a real question for businesses: if the internet increasingly distinguishes between original and generated material, what value will be placed on content that can clearly be traced back to real people, real expertise and real experiences?

AI should amplify human thought, not replace it

None of this is an argument against using AI.

In fact, I think the opposite is true. I use it too much to say otherwise, however I have set myself limits so I don’t get drunk on my own supply.

AI is an extraordinary tool for taking an original idea and helping someone develop it. It can question your assumptions, research supporting evidence, restructure your thinking, improve your grammar, help with design and make it easier to communicate an idea to the world.

But there is an important distinction between using AI to amplify an idea and using AI to manufacture one.

A business owner explaining why they started their company and then using AI to help communicate that story is very different from asking a model to “write an authentic About Us page”.

A project engineer describing what actually happened on a difficult project and having AI help turn those notes into a case study is different from asking AI to invent a polished project story from a handful of bullet points.

The technology can do both.

The question is whether it should.

As intelligence becomes cheaper and more widely available, simply producing content will no longer be particularly valuable. Anyone will be able to generate a polished website, article, image or campaign almost instantly.

The scarce resource will be having something worth saying in the first place.

And in an internet increasingly filled with generated intelligence, genuinely human ideas may become one of the most valuable things we have.