Brands seeking to elevate their visibility are producing content at an accelerated pace. The growing influence of earned media, optimized websites and thought leadership on generative artificial intelligence (GenAI) search engines is driving an uptick in content production. For many, the stakes have never been higher in the race to secure respectable answer engine optimization (AEO) and generative engine optimization (GEO) rankings. At the same time, there is no shortage of opinions on whether content should be human-generated, large language model (LLM)-generated or a combination of the two.
The ‘AI police’ are out in full force, consistently citing brands for using notorious AI buzzwords and common AI frameworks in their content. This has human writers resorting to desperate measures—intentionally planting typos into text and avoiding em dashes at all costs. Never mind that em dashes can be traced back to the 17th century; their reputation shattered when LLMs started spitting them out left and right.
Sadly, this game of ‘AI-gotcha’ has many readers so focused on the origin of the content that they gloss over the most important element of a well-written piece—the message.
The Message is Getting Lost
There are varying schools of thought when it comes to using AI. Some content creators cannot fathom generating written content without the help of their friend, Claude. And many do so loudly and proudly, leaning indiscriminately on hackneyed constructions such as ‘It’s not just X, it’s Y.” Many crank out AI slop left and right, without even bothering to remove the default Segoe UI font or the score lines after each section of an AI-generated outline.
Others use AI, but work diligently to hide their tracks, replacing em dashes with semicolons and making arbitrary edits in an attempt to throw off the AI detectors. Ironically, some of these same writers don’t think twice about defaulting to other overused AI sentence structures, such as ‘The businesses that will win’ and ‘Here’s the thing’ or ‘Here’s what everyone misses.’ Em dashes or not, they’re not fooling anyone. But, arguably, that isn’t their biggest vulnerability.
What perhaps warrants greater attention is that all the focus on supposed AI fingerprints distracts writers and readers alike from evaluating the actual substance of the content itself.
The Branding is Getting Diluted
While the AI police crack down on ‘AI slop’ and writers feverishly wordsmith AI-generated text to sound more human, the key takeaways get lost in the crossfire. Yes, punctuation matters, and yes, edgy-sounding headlines capture attention, but if the brand’s core message is drowning in a sea of AI jargon or surrounded by arbitrarily modified wording, is the content really doing its job? Sure, a brand’s messaging might surface in the AEO or GEO search engines, but if the content sounds exactly like a competitor’s messaging, it could confuse or underwhelm readers. All that branding work, credibility and differentiation can get lost along the way.
This is not to say that using AI to produce content is unequivocally bad. LLMs can be useful for brainstorming, surfacing sources or historical references, producing rough drafts—and even writing. If AI helps a brand produce content more efficiently, then more power to them. But efficiency should not be mistaken for effectiveness.
Using AI as a crutch instead of a tool can backfire, whether or not there is evidence of blatant AI use. The energy and effort content producers spend either taunting or trying to evade the AI police might be better spent vetting the messaging to ensure it is contextually sound and aligned with the brand’s known attributes before anything gets published.
Judgment is the Secret Sauce
With 41% of employers expecting to reduce their workforces as AI automates certain tasks, companies will naturally explore whether it could effectively replace their human writers. Until it can read a room, tap into human emotion, recognize subtle nuances, detect sarcasm and distinguish hallucinations from facts, human oversight should be non-negotiable. This is especially true for industries such as public relations, where reputations are on the line.
Human writers and editors can help improve AI-generated content by asking thoughtful questions like: Does the LLM understand the play on words? Is there an innuendo that is hidden from an LLM’s frame of reference? Is an inside joke about to be leaked? A breakdown in any of these areas could tarnish or destroy a well-established brand overnight.
Bad Content is a Lawsuit Waiting to Happen
AI can recognize patterns and produce remarkably human-sounding language, but it still needs human context, judgment and accountability to help ensure the messaging is on point.
A recent study published by Saturn, an AI technology firm serving the financial industry, indicated that the average accuracy rate across multiple free and purchased AI LLM models was only 43%. That margin of error is too important to dismiss.
Whether or not an LLM is involved, content should be edited, thoroughly reviewed and fact-checked. When vetting sources for credibility, writers and editors should verify publication dates and ensure studies reflect findings relevant to the intended industry and the target country. They should confirm that all evidence directly supports the stated claims.
Plagiarism is another concern. Try telling an LLM the content it produced sounds familiar. If it agrees, ask why the content sounds familiar. Don’t be surprised when it replies that it was trained on messages written by others, so of course it sounds familiar.
Failure to cross-check AI-generated content can derail an otherwise solid product launch. Worse, it can have severe consequences, including trademark violations, irreparable reputational damage and even litigation.
The risks of not properly vetting content far outweigh the added work of due diligence.
Professional Writers Keep LLMs Honest
Good content should be accurate, distinctive, useful and true to the brand. Whether AI played a role in getting it there should be secondary.
AI can help writers brainstorm, research, organize ideas and produce a first draft. The writer must still decide what belongs in the piece, what doesn’t and whether the finished product is likely to resonate with the desired audiences.
The AI police may be looking for telltale words, sentence structures and punctuation. Good writers serve a more significant role—ensuring the message serves its intended purpose and connects with readers.
Jill Casey Pintor is a financial content writer at StreetCred PR, a public relations agency focused on the financial services sector.