Gnosis: The Type of Knowledge AI Will Never Have

Turns out—a lot.

In the months since ChatGPT became publicly available, I put it through the wringer in several use cases.

We’ve had it write a section of an article, including the Clearscope recommendations in the prompt, several copy-pasted sources, and instructions for tone and style.

We even tried using ChatGPT for content review, asking it to “read” a piece of text and give recommendations to improve engagement.

Were the results perfect?

No, it was not as good a job as a professional content writer or editor would have done. But it was very close, and that was only GPT3. GPT4 is already out. GPT5 will come after it.

Many are right to ask the question if content marketing will be fully automated in the near future. If these generative AI tools can write and edit, surely they can also perform automatic keyword research and create content briefs (that they would then fill out themselves).

Why can AI do all of this?

In the simplest terms possible, tools like ChatGPT use the technology of Large Language Models (LLM) and the vast amounts of data they’re trained on to give the most statistically probable answer to a prompt.

LLM

A large language model is a type of artificial intelligence (AI) algorithm that uses deep learning techniques and massively large data sets to understand, summarize, generate and predict new content.

That’s a significant oversimplification, by the way, but it will work for this article.

This means that AI can do what it does because it can access many kinds of data. Contemporary and classic literature, blogs, news sites, scientific works, and social media, to name a few.

But is something missing? (Aside from in ChatGPT’s case, data on current events past September 2021).

To find out what AI is missing, we first need to go back by quite a bit. All the way back to ancient Greek times, to be exact.

And we need to remember something that should be obvious—that all this talk about AI and content generation is primarily to do with knowledge.

So, with those things in mind, here’s how the Greek philosophers classified knowledge:

• Doxa: hearsay or received knowledge, the things we hear from others such as family, friends, and teachers
• Episteme: what we most often think of as knowledge. It comes from observation, measurement, experimentation, and reasoning.
• Gnosis: the knowledge we get from having experienced something

How does this relate to AI and LLMs? Well, because of the dataset they were trained on, tools like ChatGPT are masters of doxa. In a very literal sense, they have “heard it all.”

Among the things they “heard” are scientific facts, results of studies, and experiments, so we could say that AI also has access to the episteme of the entire world.

What it doesn’t have access to is direct human experience and perception. It’s doubtful that it ever will unless some kind of cyberpunk Neuralink-for-all scenario becomes a reality.

For now, our experiences and original opinions remain our own.

What does all that have to do with content marketing? Glad you asked.

If AI is poised to replace humans in the kind of content that most agencies produce today, we also know what type of knowledge it doesn’t possess. There’s a clear implication: use our unique knowledge to create original content.

That’s not to say that “classic” articles, blogs, and lists will disappear. They won’t; generative AI tools will just produce them.

And make no mistake; this content will be good, and that’s why the AI takeover is scary in the first place.

But, companies that’ll want to differentiate themselves from this perfect content deluge will want something new.

This new thing can be content based on human experience—for instance, articles with original research based on user interviews.

Or interview a company’s subject matter experts to produce content that will help them establish themselves as thought leaders in their space.

Essentially, content marketing might have to take a page from the journalism book.

This means making content from unique sources with an interesting angle or a twist. It means “invented concepts” and “contrarian content” that will provoke people and get them to think.

There is a benefit to this for the clients of content marketing agencies.

Pursuing this “human-centric” content means a company can say, “Look at us! We don’t have the same guides and tutorials as everyone else. We’re knowledgeable and opinionated about our niche, and our experts want to share that knowledge with the world.”

Because remember, AI content is already here, and it will take over more and more, and it will be good. But if everyone has that good content, how do you stand out?

By focusing on the strengths of human beings that AIs don’t possess.

It’s not a secret that ChatGPT and other generative AI tools have made many creatives and marketers get in a tizzy.

But, there’s reason to be cautiously optimistic. AI does have the potential to liberate us from the menial and boring while unlocking a whole new world of creativity.

In this world, however, there will still be room for compelling, engaging, and interesting stories that only humans can produce.