The argument against AI in creative work is not new, but a blog post from Slice Communications puts it more plainly than most: AI produces stolen, regurgitated content — the word "content" itself being, in the author's view, the perfect label for filler that devalues filmmaking, graphic design and audio production.
The author admits to being a self-described curmudgeon and skeptic. The concern is not just aesthetic. AI copy, the argument goes, cites no original sources, credits no original thinkers and produces what amounts to mashed-up years-old text. The internet already has plenty of useless copy. A dull post generates minimal engagement. If the assignment doesn't inspire anything worth saying, the better answer is to rethink the assignment — not hand it to a machine.
The practical case colleagues make for AI, according to the post, runs like this: use it to get started when ideas run dry, use it for outlines, apply it to copy where minimal creative input is required. The author finds this unconvincing. If a piece of writing needs AI to exist at all, the question worth asking is whether it needs to exist.
There is one narrow concession: AI works as a brainstorming tool, a springboard toward better ideas. The author grants that much. But the post is explicit that AI has never produced something the author found truly surprising or insightful — and that surprise and insight are not, in any case, what AI is optimized for.
To test the machine's capacity for self-advocacy, the author asked ChatGPT to write a blog post for a marketing agency defending AI to a skeptic. The result is instructive. ChatGPT opened by defining artificial intelligence as "computer systems that can perform tasks that typically require human intelligence, such as speech recognition, decision-making, and language translation" — a sentence that is accurate, dry and completely unremarkable. It then listed machine learning, natural language processing and computer vision as types of AI, and moved to its first tip: "Embrace the fact that AI is here to stay."
The irony is intentional on the author's part. ChatGPT, asked to make the creative case for AI, produced exactly the kind of competent, bloodless copy the post argues AI always produces. No surprising angle. No specific number. No named person. Just the structural outline of an argument, dressed in the language of a corporate FAQ.
The business model question here is not trivial. Marketing agencies bill for creative output. If AI compresses the time required to produce a first draft from four hours to 20 minutes, the agency either charges less, cuts headcount or both. Clients who understand this will push for lower rates. Agencies that resist will lose price-sensitive accounts. The ones that thrive will be the ones that use AI on commodity tasks — event-date posts, regulatory disclosures, routine social copy — while protecting billing hours for work that actually requires a human point of view.
The content calendar problem is real. Marketing teams face deadlines regardless of whether they have anything worth saying. The post acknowledges this directly: the "siren song" of AI is loudest when a writer is burned out and the well is dry. That is also when AI does the most damage — filling a calendar slot with text that no one reads and no algorithm rewards.
Search engines increasingly penalize thin, repetitive content. Google's Helpful Content updates, running since 2022, specifically target pages that exist to fill space rather than answer a question. An AI-generated post that restates widely known facts without a named source, a specific number or an original observation is exactly the kind of content those updates are designed to suppress. The agency that floods its blog with ChatGPT output to hit a posting frequency is not building an audience — it is building a liability.
What the post does not say, but the business logic implies: AI's real value in marketing is in tasks that don't require a voice. Data extraction from campaign reports. A/B test variant generation. Translating one approved piece of copy into five language markets. Drafting a legal disclaimer from a template. None of that is what marketing agencies lead with in a pitch, but all of it compounds into real hours saved.
The author ends on a rallying cry for poets and a dig at anyone who mistakes volume for quality. The ChatGPT excerpt the post includes does not help AI's cause — it is, functionally, the proof of the argument against it. AI is here, the haters are correct about what it produces at its worst, and the agencies that figure out the narrow set of tasks where it earns its keep will outcompete the ones still debating whether to use it at all.
