Generative AI adoption has reached over 50% in three years, and the tells are everywhere. Most of its content is easily identifiable because teams draft in ChatGPT or Claude, make light edits, and hit publish. Customers and stakeholders are learning how to spot such manufactured communication, and the jury is out on its effect on truth, accuracy and persuasion.
Here is what gives AI writing away:
- The “It’s Not X, It’s Y” tic: The most reliable reveal is negative parallelism. “It’s not about the prompt, it’s about the context.” “This isn’t a feature update. It’s a category shift.” Every AI model does this multiple times per output. Research shows large language models (LLMs) gravitate toward the same grammatical structures in a way that just isn’t human. Real writers occasionally use this structure for emphasis, but AI overuses it to make a shallow idea sound profound. Using it even once is enough to make a reader question who actually wrote the piece.
- The word list that outs every AI draft. If you’ve ever skimmed through an article and noticed the writer “underscores” their point, “showcases” examples, or “delves” into the history, it’s a dead giveaway they’ve used AI. Such verbs aren’t wrong, just unnaturally mechanized and overused. A LinkedIn post with three or four of these predictable constructions reads as machine to anyone judging authenticity. Tools like GPTZero can scan for these patterns before your readers do. It’s machine against machine.
- The rule of three. Good writing is clear, concise, and compelling. You’ve read that exact sentence before. Everyone has. LLMs group three adjectives because they’re baked into the communication strategies they trained on. A quick scan through any bylined article or executive post for three-part constructions reveals where AI did the heavy lifting and where a human edit is needed.
- When AI makes up the facts. Ask Claude or ChatGPT for a statistic, and it will produce one that sounds right, whether its source is properly interpreted or actually exists at all. The threat of so-called hallucinations is a grave danger to any communication that appears under a human name or company. The fix is suspicion—and the challenging of one AI statistic by a similar prompt to a sister artificial intelligence source. AI’s people-pleasing instinct causes happy numbers designed to make us smile. A new chat, ordered to play the nitpicking auditor, breaks the loop and makes AI find errors instead of confirming them.
- Punctuation overload. Em dashes aren’t new, but there’s a current epidemic of them in generated content. If an op-ed or other persuasion document looks like Morse code, increase skepticism. The formal colon, once rare in information writing, is another giveaway: inserted before every observation as if it needs a drumroll. Over-frequent punctuation, apparently taught to large language models by 1950s schoolmarms, becomes a useful tell. Human editors recognize such awkwardly formalized patterns immediately. Be one!
We have in our hands a new and marvelous tool for first drafts, but the decision to publish still requires our head. To err is human, and despite the charm of our new artificially intelligent assistant, mistakes are still on us.
OWC is here to help keep it real.
IN THE NEWS
California Virtual Academies (CAVA) student featured in the Los Angeles Times
The Los Angeles Times featured CAVA student and K12 Talent Showdown finalist Francesco. The story highlights Francesco’s life as a fifth-generation circus performer traveling throughout California with his family’s Caballero Circus while balancing school and performances. Read the full story here.
L.A. Sentinel spotlights Crenshaw Family Health Center groundbreaking
OWC secured coverage in the Los Angeles Sentinel highlighting the groundbreaking of Saban Community Clinic’s new Crenshaw Family Health Center, the clinic’s expansion into South Los Angeles. The new facility will provide comprehensive medical, dental, behavioral health, and support services for thousands of patients. Read the full story here.
23ABC Bakersfield featured iQ Academy (iQLA) student on graduation and future career
23ABC Bakersfield highlighted iQLA student Samuel, who graduated at just 16 years old with a 4.0 GPA and college credits already completed. Samuel overcame academic challenges, set a goal to graduate early, and used online learning and dual enrollment opportunities to accelerate his education while pursuing his dream career. Watch the full story here.
ICYMI: Talk PR with Christian Williams
The latest episode of Talk PR featured former Washington Post editor and television writer/producer Christian Williams. Williams discussed how he built an online following with more than 12 million YouTube views and uses original content as the sole marketing tool for his successful books. He shared insights and best practices for businesses looking to strengthen their digital storytelling and audience engagement. Tune in here.
