For our 25th set of trends and predictions for 2026, we said AI slop–defined as “low-quality, AI-generated content”–will generate a backlash. We cited a Wall Street Journal article that reported that companies are seeking storytellers.
This week, Business Insider reported on the same trend, “The Hottest New Job in Tech is Writing Words.” According to BI, demand for human writers is booming because AI-generated content tends to be soulless, generic, unnuanced and uninteresting. (Please note: This blog–despite any em-dashes–was written by humans only; we do use Grammarly to help proofread before publishing. We mention this because Grammarly helps us catch missing words when we’re typing too fast, for example, and because we’re not luddites.)
According to BI, writers are being hired to use AI because a human writer can improve AI content so it doesn’t read like slop.
In any case, it validates our prediction that people don’t respond well to slop. One of benefit for companies hiring writers to develop content is that if unedited, AI-generated content is not copyrightable. Organizations that use AI-generated content as is can’t protect it, which means a competitor could use the same language on its website. But heavily edited/re-written content probably can be copyrighted. So there’s that. (We’re not lawyers, so check with an attorney.)
By the way, we’ve tested several AI detectors to assess their accuracy. In some cases, these apps accurately spot AI-generated content. But, as a test, we uploaded non-AI content to see what the detectors would find. In several cases, those pieces were scored 25% or higher for possible AI content–even when the piece was entirely written by humans. That includes one piece written and published before ChatGPT, so we know it was written entirely by a real person.
When we asked ChatGPT and Claude why some human-generated work was flagged as possibly containing AI content, they said that humans who write too smoothly are likely to be flagged as AI content. Or sometimes it’s a common phrase that gets detected. In the case of this blog, one detector said this was human-written–but identified the following sentences as possibly AI-generated: “So there’s that. (We’re not lawyers, so check with an attorney.)”
The important thing is to avoid producing slop of any kind. People can tell, and they don’t like it.
We will continue to track how our predictions are doing. Please let us know if you agree or have different insights.
