AI sameness: It's not the use of AI, but how it used

While linguistic patterns do exist and certain words have increased in frequency, the real issue is not the presence of AI, it is the absence of a distinctive voice (Source: @ Novelty Insider Novelty Insider

The idea that audiences or professionals can consistently “spot the AI” does not hold up under scrutiny”. Gavin Moffat co-founder of join.the.dots says, “Detection is far less reliable than assumed.”

He adds that audience priorities are also different from what many expect.

“Outside of high-trust environments such as journalism, readers are not primarily concerned with how content is created.

“Their focus is on whether it is useful, clear, and worth their time. Utility and relevance consistently outweigh origin.”

“AI fingerprints”

Certain words, phrases, and patterns such as “quietly,” “silently,” and “In a world of…” and how overly polished sentence structures have become increasingly common, creating a sense that much of today’s content sounds the same.

The concern is that if everything starts to feel algorithmic, brand voice and credibility are at risk.

The original conversation focused on a growing frustration among communication professionals.

This follows recent industry discussions on so-called “AI fingerprints” in business writing, as new evidence-based data is challenging widely held assumptions about how AI-generated content is identified and, more importantly, what matters to audiences.

The AI fingerprint discussion

While linguistic patterns do exist and certain words have increased in frequency, the real issue is not the presence of AI, it is the absence of a distinctive voice.

When content lacks clarity of thinking, originality, and brand-specific tone, it becomes interchangeable, regardless of whether it was written by a human or generated by a model.

How AI is used

The issue of AI sameness is not caused by the use of AI itself, but by how it is used. Many content processes stop too early. Teams generate content, review it superficially, and publish without applying the necessary layers of refinement.

The result is a predictable structure, generic language, and a loss of differentiation.

The evidence-based findings reinforce a simple but critical principle - Generic input produces generic output.

When organisations introduce brand context, apply clear style guidelines, and include a deliberate human review process, the detectability of AI-driven patterns drops significantly.

More importantly, the content begins to sound like the brand again.

“The central risk facing organisations today is not being exposed for using AI but rather sounding exactly like everyone else,” says Moffat.

What these findings mean for the future

Originality economy to a distribution economy

Brands spending time trying to “sound less AI” without improving the thinking behind the content are optimising for the wrong metric. The shift is already underway.

We are moving from an originality economy to a distribution economy, where advantage lies not in who writes first, but in who frames, contextualises, and delivers value most effectively.

Brand voice a measurable asset

Brand voice is no longer a soft concept. It is becoming a measurable asset. As hybrid content continues to outperform, voice, tone, and consistency will be tied directly to engagement, memorability, and performance.

Style guides will move from brand wallpaper to operational tools, and AI workflows will need to preserve tone, not just produce volume.

Reshaping content - again

How content is produced is reshaping again. The role of content production will split between those who generate and those who refine.

The value will sit with those who shape meaning, not those who produce first drafts.

Trust is shifting

It is no longer anchored in authorship, but in consistency. Audiences may not detect AI, and in most cases, they do not care. What they do notice is whether a brand sounds like itself, delivers value, and shows up coherently over time.

The hybrid approach

“The fastest shift will be in content production roles splitting in two,” says Moffat.

“Not because it is the most strategic move, but because it is the most painful problem to ignore.

“We’re seeing content volume surge and as more first drafts become cheap, teams are finding themselves producing more while liking less of what they produce.”

He says that the most effective content is not purely human or purely AI-generated.

“It is a hybrid approach that combines AI-driven efficiency with human judgement, tone, nuance, and contextual understanding.”

He adds, “The slowest shift, it likely to be the move from originality to distribution and framing advantage, this is because it requires strategic maturity.”