You have seen it this week. The LinkedIn post that opens with "In today's fast-paced world". The product image where the hands have six fingers. The carousel that says a lot and means nothing. Most people can now spot AI content within seconds, and more of them are scrolling straight past it.
Here is our position, and we expect some of you to disagree: AI is not the problem. Lazy AI is. Businesses are right to use AI to produce more, faster and cheaper. But when volume becomes the strategy, the output turns into spam, and spam costs you the one thing you were trying to buy: attention.
We know this because we made the mistake ourselves.
The numbers: AI content is now half the web
This is not a feeling. The data shows how quickly AI content has taken over.
- Half of new articles are AI-written. SEO firm Graphite analysed English-language articles published on the web. Within a year of ChatGPT's launch, 39% were AI-generated. In November 2024, AI-written articles overtook human-written ones, and the share has hovered around half since. (Graphite)
- More AI content does not mean more reach. Graphite also found that these AI articles largely do not appear in Google or ChatGPT results. Publishers flooded the web, and the flood mostly went unseen.
- Advertisers are all in. 83% of US ad executives say their company now uses AI in creative work, up from 60% in 2024 (IAB, January 2026).
- The public has a name for it. Merriam-Webster made "slop" its 2025 Word of the Year, defining it as digital content of low quality that is produced usually in quantity by means of AI (TODAY).
In other words, companies have scaled production much faster than audiences have scaled their patience.
"I can spot it from a mile away"
The audience has noticed, and it is not impressed. Social listening firm Meltwater counted about 2.4 million online mentions of "AI slop" in 2025 up to 20 November. That is nine times the same period in 2024, and negative sentiment peaked at 54% in October 2025 (Meltwater).
Scroll through any comment section and you will see the same complaints:
- The look: waxy skin, glossy lighting, extra fingers, text that melts into gibberish.
- The voice: "Unlock", "elevate", "game-changer", "in today's fast-paced world".
- The sameness: ten brands posting the same carousel, with the same hook, in the same week.
- The feeling of being tricked: a brand that looks like it could not be bothered to talk to you.
The gap between what marketers believe and what audiences feel is widening. In IAB research published in January 2026, 82% of ad executives thought Gen Z and millennial consumers felt positive about AI-generated ads. Only 45% of those consumers did. Consumers were also twice as likely as executives to call brands that use AI in ads "manipulative" (IAB).
Big budgets do not protect you either. In December 2025, McDonald's Netherlands pulled an AI-generated Christmas ad within days after viewers called it "creepy" and "inauthentic" (NBC News). Platforms are responding too. On 15 July 2025, YouTube tightened its monetisation rules to target mass-produced, repetitive content (Search Engine Journal).
Why businesses keep doing it anyway
It would be easy to wag a finger here. We won't, because the business case for AI content is real.
A small business that once paid a designer, a copywriter and a video editor can now produce a week of posts in an afternoon. A marketing team of two can test ten ad variations instead of one. For a South African SME watching every rand, that is not a gimmick. It is survival. The IAB found that 86% of video ad buyers are using or planning to use generative AI for creative work (IAB). Nobody is going back.
The trap is that output is easy to measure and attention is not. You can count posts published. It is harder to count the followers who quietly stopped reading because every post sounded like the last one. So companies optimise for what they can see, and the cost of the rest builds up in the background.
Our confession: we made the slop too
We are an AI consultancy, so we tried the tools first and hardest. And for a while, we produced exactly the kind of content this article complains about.
The pattern was simple. A one-line prompt went into a writing assistant, and a polished paragraph came out. A single image went into a free video generator, and a cinematic clip came out. It felt productive. It looked professional. But when we lined the posts up side by side, they could have come from any company in any country. The headlines in our AI videos warped. The copy had our name on it but not our voice.
The lesson cost us time and reach, and it was not that AI is bad. It was that a tool will only give you back as much thinking as you put in. So we went back to the drawing board. We stopped asking AI to create the content and started asking it to help us build it. Human strategy came first, then AI did the heavy lifting on drafts, variations and editing.
That way of working is now built into our free Prompt Builder. It walks you through the CRAFT method: Context, Role, Action, Format and Tone. It sounds basic, but most generic output comes from skipping the first two steps.
What using AI well actually looks like
The fix is not less AI. It is AI with a person in charge. These are the rules we now work by:
- Start with something only you know. A client question you answered last week, a mistake you made, a local detail. AI cannot invent your experience, so give it that experience to work with.
- Give the tool context, not just a command. "Write a post about invoicing" gets you slop. "Write for a Joburg events planner who chases late payments on WhatsApp" gets you a draft worth editing.
- Treat the first output as a draft. Cut the filler words, add an opinion, and read it aloud. If it could have come from a competitor, rewrite it.
- Keep text out of AI images and video. Generate the visual, then add headlines and branding in an editor. Warped text is one of the fastest giveaways.
- Publish less, but better. One post that starts a conversation beats seven that nobody remembers.
- Measure attention, not output. Track saves, replies and enquiries, not how many posts went out.
Get it right the first time
We learnt all this the hard way so that you don't have to. At Okiru, we help businesses and creators choose the right AI tools, set up workflows that keep their voice, and train their teams to use AI with judgement. The tools handle the volume. Your people keep responsibility for the strategy, the story and the results.
Two free places to start on okiru.biz: the AI Tool Advisor, a searchable directory of vetted AI tools, and the Prompt Builder. If you want hands-on help, our AI training and advisory team can work with you directly.
Sources
- Graphite: More articles are now created by AI than humans (October 2025)
- IAB: The AI Ad Gap Widens (15 January 2026)
- Meltwater: What the rise of AI slop means for marketers (2025)
- TODAY: Merriam-Webster's 2025 Word of the Year is "slop" (15 December 2025)
- NBC News: McDonald's pulls AI-generated Christmas ad (December 2025)
- Search Engine Journal: YouTube targets mass-produced content (July 2025)
Okiru. Smart data, smarter transformation.