Content Differentiation in the Age of AI

As quickly as we saw think pieces extolling the virtues of letting AI run the content marketing, we saw rebuttals about the infiltration of “AI slop.” So with all of AI’s generative capabilities, why would we bother writing our own content?

Put simply, it’s because at Proven, we operate with humans in the lead, not merely in the loop. For instance, this Insight was crafted without the use of AI (as were its images). That may surprise you given our AI Native, Human Led philosophy, but this culture lets us decide when the benefits of a human touch outweigh speed or volume. But even when AI plays a role, a human must have the final say.

Whether you use AI to help outline, edit, or defeat the intimidation of a blank page, these tips can help ensure your content is furthering your goals, not getting written off as AI.

Defining “Slop”

AI slop isn’t defined by the model that produced it, though some models are better “writers” than others. It’s not any of the “tell-tale signs” that give AI content away, it’s the lack of substance. Slop tends to be lengthy without actually saying much of anything. It hedges so there’s nothing to disagree with; it makes claims without intending to back them up; and it lacks personal flair, pretending to offer an anecdote so general it’s meaningless.

If slop happens when there’s no substance or point of view, a key to combating it is to start with what only you know.

Highlight Your Point of View

For all its generative capability, AI doesn’t have its own point of view. Whether it’s ChatGPT, Claude, Gemini, or another LLM, these models are based on prediction. They’re trained to supply what is most likely to come next in an answer or a paragraph. Give a generic prompt and you’ll get a generic answer.

This is where human discernment and creativity cannot be replaced. AI can tell you what’s generally said about a topic, but only you can tell readers about your perspective. You know the project timeline that was saved by an outside-the-box idea, you remember the time that you ran into a client at a conference and turned it into a social media campaign, and you tested the new tool for a UX exercise. Your point of view is crucial to creating content that connects.

What’s the Harm in Seeing What AI Says First?

To share my personal point of view, I double majored in marketing and economics in undergrad, and one of my favorite classes was Introduction to Behavioral Economics. Here, I learned about the Anchoring Effect, first introduced by Amos Tversky and Daniel Kahneman. They found that people’s decisions are unconsciously influenced by a reference point or “anchor,” regardless of its accuracy or applicability. This cognitive bias has been tested in multiple studies, across a number of scenarios and still holds.

The Anchoring Effect is why I think the inclination to “see what AI says first” is a slippery slope. If you approach an LLM with little more than a general topic, you may unconsciously let it influence you and take the model’s output as your own opinion. Once your perspective is already on the page, then you can bring AI in to help.

Don’t Get Too Hung Up On the “Tell-Tale Signs” of AI

Generative engines were trained on centuries of writers: how they formulate arguments, structure sentences, and phrase cadence. By design, these models lean on facets of writing that many of us use. It’s no wonder that things in your authorial toolbox (or pencil case) get called out as a “tell-tale sign” of AI, whether you used it or not.

The Rule of Three

One supposed signal of AI is simply the Rule of Three at work. In storytelling, the Story Triangle encompasses the Storyteller, the Audience, and the Story itself. There’s the three-act structure of the Hero’s Journey: Departure, Initiation, and Return. In his book Screen Adaptation, Eric R. Williams uses a device called the Triangle of Knowledge to build tension by depriving one of the three points on the triangle (protagonist, second character, or audience) of knowledge that the other two share.

From slogans (Reduce, Reuse, Recycle) to iconic oratory (Friends, Romans, Countrymen…), the Rule of Three is alive and well and has been for centuries. To eliminate it because the generative models employ it liberally robs your writing of a strong literary device.

Justice for Em Dashes

I love an em dash. For me, they allow me to write the way my internal monologue speaks, and many writers love them for similar reasons. This natural speech pattern is also why so many AI models use them, precisely because it sounds more human. You can certainly write your way around an em dash if you’re so inclined, but if the writing is stronger with it, just let it be.

Purposeful Typos

An interesting tactic making its rounds is including purposeful typos to add an obvious nod to human nature and fallibility. It emphasizes the lengths some are willing to go to in order to signal human authorship and avoid side-eyed AI suspicion.

In our industry and in our client work, however, this is more likely to signal haste, lack of oversight, or poor skill. If you’re writing as your personal brand, it’s less risky a choice than if you’re writing on behalf of others, but in my personal writing all typos are accidental, not purposeful.

Rather than planting mistakes, our preference is to edit like a human, for humans.

Edit Like a Human

No matter what role AI plays in the draft of a piece, a human edit is the absolute bare minimum. Verify every fact, statistic, and citation in the piece to ensure hallucinations are caught and nuance isn’t over-simplified.

Next, read the whole thing aloud. Great content has a natural rhythm, and your ear is the best editor. Sentences that your eyes skim past may sound awkward and stilted when read off, or they may even sound like something you’d never actually say. Edits here are crucial to ensuring your piece feels human.

Worth Saying, Worth Remembering

It’s not my intention to have you swear off AI — its use and use cases are things you must decide for yourself, in accordance with your organization’s policies. However, if you do use it, it’s your responsibility to ensure that whatever you publish is both accurate and grounded in a point of view worth sharing.

Readers won’t remember if you used an em dash, made a typo, or listed three examples, but they will remember if what you wrote was worth their time.