By Brian French, Florida Authority Network — September 15, 2026
Quick Answer: Most AI content deserves the “slop” label. But a third category, “good enough” content that tells a story with 85–90% accuracy at a fraction of the cost, is the one that should worry professional creators. Casual readers don’t grade on a Pulitzer curve, and AI keeps getting better at hitting the bar they actually use.
Why “slop” became the word of the year
“Slop” was not a niche complaint by the end of 2025. Macquarie Dictionary, Merriam-Webster, and the American Dialect Society all named it Word of the Year, reflecting broad unease with the flood of low-grade AI output online. Meltwater tracked a ninefold increase in mentions of “AI slop” from 2024 to 2025, with negative sentiment peaking at 54 percent in October.
The volume behind that reaction is real. Graphite analyzed 65,000 English-language URLs from Common Crawl and found that by November 2024, 50.3% of new web articles were primarily AI-generated, up from roughly 5% before ChatGPT launched.
So the scorn is earned. The mistake is assuming every piece of AI content sits in the same bucket.
The three categories of AI content
| Category | What it looks like | Who it hurts |
|---|---|---|
| The Bad | Generic, bloated, written for algorithms | Readers, platforms |
| The Ugly | Wrong facts, fake sources, hallucinated quotes | Everyone, including trust in news |
| The Good Enough | Readable, mostly accurate, cheap, fast | Professional creators |
The Bad: content nobody asked for
This is the bulk of the slop. It is mass-produced to satisfy algorithms rather than help real readers, and its spread damages trust and makes it harder for quality work to stand out. The economics explain the volume: near-zero production cost meets advertising revenue that is indifferent to quality, with AI book farms churning out 10,000 to 40,000 Kindle e-books a month.
The search engines have noticed. Graphite’s companion study found that 86% of articles ranking in Google Search are human-written and only 14% are AI-generated, and that 82% of articles cited by ChatGPT and Perplexity are written by humans. Bad content gets made in huge quantities and then mostly gets ignored.
The Ugly: confident misinformation
The Ugly is worse than the Bad because it looks credible. An EBU-coordinated study led by the BBC, covering 22 public broadcasters in 18 countries, evaluated over 3,000 AI assistant responses and found 45% had at least one significant issue, 31% showed serious sourcing problems, and 20% contained major accuracy errors such as hallucinated details or outdated information. Examples ranged from falsely stating surrogacy is illegal in Czechia to naming Pope Francis as the sitting pope a month after his death.
That matters for the news business specifically. The Reuters Institute’s 2025 Digital News Report found 7% of online news consumers already use AI assistants for news, rising to 15% of under-25s. Ugly content is a real problem, and it is the category that makes the purist’s case for them.
The Good Enough: the category that should worry you
Here is the uncomfortable part. Strip out the Bad and the Ugly, and what remains is content that is coherent, reasonably accurate, and pleasant to read. It is not brilliant. It does not need to be.
The research on reader perception is blunt. Villanova University researchers had 1,682 people read one of six short stories, three by published human authors and three generated by ChatGPT. Readers rated the AI stories as higher quality and more absorbing than the human ones, and follow-up experiments showed people guessed the author correctly only 40% to 52% of the time, no better than chance. The researchers noted the AI versions tended to state their themes explicitly rather than making readers infer them, and suggested people may simply prefer predictability because subtle material takes more effort.
That is the whole story of Good Enough in one finding. Casual readers are not looking for Ken Burns. They want the story told clearly, quickly, and mostly right.
Why 85–90% accuracy wins
Consider what “accuracy” means to different audiences:
- A litigator, an auditor, or an academic needs 99%+ and will pay for it.
- A business owner scanning industry news at 6 a.m. needs the gist, the names, the numbers, and the direction of the story.
- A consumer reading about a new restaurant, a hurricane forecast, or a zoning fight needs the picture, not the footnotes.
The second and third groups are the overwhelming majority of readers. For them, content that is 85–90% accurate, delivered today, at pennies per piece, beats content that is 99% accurate, delivered next week, at hundreds of dollars per piece. Or behind a pay wall. Not on merit. On value.
Graphite’s researchers explicitly noted that they did not evaluate AI-assisted content with heavy human editing, and that they believe it may be an effective strategy. That hybrid is where Good Enough lives, and it is the category no detection tool or dictionary is tracking.
What the purist gets right, and wrong
The purist is right that voice, judgment, and original reporting still matter. Audiences are not anti-technology so much as anti-blandness; they want opinion, specificity, and visible human effort, the qualities AI defaults away from. Graphite’s own read is that the market is becoming a symbiosis rather than a dichotomy, with the most effective publishers integrating AI into human workflows rather than replacing them.
The purist is wrong that readers will pay a premium for the last 10% of polish on ordinary stories. University of Florida economic modeling found that low-quality AI content hurts both consumers and professional creators, but that if AI quality improves enough, it could ultimately help consumers, novice creators, and professionals alike. The bar keeps rising. The cost keeps falling. Good Enough gets better every quarter.
Brian’s Take
I ran money for a living before I ran a digital newsroom, and this is a cost-of-capital problem dressed up as a taste argument. When the marginal cost of a readable, mostly accurate story approaches zero, the market clears at Good Enough. It always has. Paperbacks beat leather bindings. Digital photos beat film. Nobody won an Oscar with a phone camera, and nobody cares.
The mistake most publishers make is fighting for the wrong 10%. They pour money into perfecting prose that readers skim. The defensible ground is the part AI cannot manufacture on its own: verified local facts, a named human standing behind the analysis, proprietary data that does not exist anywhere else, and the judgment to say what a story means. That is the model we built the Florida Authority Network on: technology gathers and structures the public record, and a human analyst verifies it and adds the take.
Good Enough is not the enemy of quality. It is the new floor. The question for every content business in Florida is what you are building on top of it.
Frequently Asked Questions
Is AI content still penalized by Google?
Google penalizes scaled content made primarily for rankings, not AI use itself. Google’s March 2024 core update targeted “scaled content abuse” and reduced low-quality results by roughly 45%. Human-reviewed, original AI-assisted content is not the target.
Can readers actually tell the difference?
Mostly no. A Tooltester survey of over 1,900 Americans found only 40.8% of people unfamiliar with generative AI could correctly identify AI content, while 80.5% believe publishers should disclose when AI was involved. The disclosure expectation is the part publishers should take seriously.
Does “good enough” mean accuracy doesn’t matter?
No. It means the last few points of accuracy carry a cost most readers will not pay for on routine stories. For financial, legal, and medical reporting, the bar is higher and human verification is non-negotiable.
What should a small publisher do?
Use AI for research, structure, and first drafts. Put a real person’s name and judgment on the final product. Build something proprietary, such as local data or an index, that AI alone cannot produce.
Sources and Further Reading
- University of Florida News, “‘AI slop’ hurts consumers and creators. But high-quality AI could help both,” March 4, 2026 — https://news.ufl.edu/2026/03/ai-slop/
- Graphite (Five Percent), “More Articles Are Now Created by AI Than Humans,” October 2025 — https://graphite.io/five-percent/more-articles-are-now-created-by-ai-than-humans
- Graphite (Five Percent), “AI Content in Search & LLMs” — https://graphite.io/five-percent/ai-content-in-search-and-llms
- eWeek, “AI Now Writes Half of the Internet, but Still Ranks Behind Humans,” October 20, 2025 — https://www.eweek.com/news/ai-writes-half-internet/
- Axios, “AI-written web pages haven’t overwhelmed human-authored content, study finds,” October 14, 2025 — https://www.axios.com/2025/10/14/ai-generated-writing-humans
- European Broadcasting Union, “Largest study of its kind shows AI assistants misrepresent news content 45% of the time,” October 22, 2025 — https://www.ebu.ch/news/2025/10/ai-s-systemic-distortion-of-news-is-consistent-across-languages-and-territories-international-study-by-public-service-broadcaste
- Journalism.co.uk, “Nearly half of AI-generated answers contain an error, finds EBU-BBC report,” November 14, 2025 — https://www.journalism.co.uk/nearly-half-of-ai-generated-answers-contain-an-error-ebu-bbc-report/
- EurekAlert / Cambridge University Press, “People prefer stories written by AI – especially when told they’re written by a human,” August 4, 2026 — https://www.eurekalert.org/news-releases/1138206
- Digital Trends, “Study finds readers rate AI-written stories higher, but still trust the ‘human’ label more,” August 5, 2026 — https://www.digitaltrends.com/computing/study-finds-readers-rate-ai-written-stories-higher-but-still-trust-the-human-label-more/
- Digital Watch Observatory, “AI slop’s meteoric rise and the impact of synthetic content in 2026,” February 25, 2026 — https://dig.watch/updates/ai-slop-content-social-media
- Euronews, “2025 was the year AI slop went mainstream,” December 28, 2025 — https://euronews.com/next/2025/12/28/2025-was-the-year-ai-slop-went-mainstream-is-the-internet-ready-to-grow-up-now
- Tooltester, “Survey: ChatGPT and AI Content – Can people tell the difference?” — https://www.tooltester.com/en/blog/chatgpt-survey-can-people-tell-the-difference/
- SlopDetector, “The Rise of AI Slop,” February 12, 2026 — https://slopdetector.org/blog/rise-of-ai-slop
About the Author
About Brian French
Led by a commitment to tech-intelligent curation, Brian French tracks and analyzes the Business News in Florida including corporate developments and breaking news defining Florida's economy. Brian brings an extensive financial background to his analysis, having graduated from the University of South Florida in Finance and serving as a Vice President and Portfolio Manager for Merrill Lynch Private Investors and the Trust Department in St. Petersburg, FL, as well as a Vice President and Trust Investment Officer for SunTrust Bank in Sarasota, FL. His writing blends macroeconomic trends, fiduciary capital markets, corporate strategy, and modern digital insights for a sophisticated look at Florida's business economy.
