THE AI SLOP PROBLEM: WHAT EVERY BUSINESS OWNER AND CONSUMER NEEDS TO KNOW
THE LOCAL AIM — DUE DILIGENCE DESK
Consumer Advisory · Business Advisory · April 2026 · Updated October 2026
Sources cited and linked. Methodology disclosed. Updated when facts change.
CORRECTION (October 2026): An earlier version of this article included a medical example, an engagement statistic, and an agency-ranking claim that did not meet this desk's own sourcing standard. They have been removed or replaced with sourced material. We hold ourselves to the same rules we apply to everyone else.
SOURCES
Merriam-Webster, Word of the Year 2025
https://www.merriam-webster.com/words-at-play/word-of-the-year
Graphite, "AI Now Writes as Many Online Articles as Humans Do"
https://graphite.io/five-percent/ai-now-writes-as-many-online-articles-as-humans-do
Meltwater, AI slop social listening analysis
https://www.meltwater.com/en/blog/ai-slop-consumer-sentiment-social-listening-analysis
The Guardian, "'Dangerous and alarming': Google removes some of its AI summaries after users' health put at risk," January 11, 2026
https://www.theguardian.com/technology/2026/jan/11/google-ai-overviews-health-guardian-investigation
Thomas Germain, BBC Future, "I hacked ChatGPT and Google's AI – and it only took 20 minutes," February 18, 2026
https://www.bbc.com/future/article/20260218-i-hacked-chatgpt-and-googles-ai-and-it-only-took-20-minutes
The confidence doesn't track the accuracy. That is the problem.
The internet is not filling up with lies. It is filling up with answers nobody checked. The difference matters — but only if you know to ask.
What AI Slop Actually Is
The term has a definition now. Merriam-Webster named "slop" its 2025 Word of the Year, describing low-quality digital content mass-produced with artificial intelligence.
It is not limited to obvious fakes. It shows up in search results. In marketing emails. In the AI Overview at the top of your Google search. In the blog post your next vendor published last Tuesday using a tool that cost them nothing and took four minutes.
*****************************************
What the numbers show:
About half — Roughly half of new English-language articles online are now mostly AI-written. Graphite measured 50.9% in late 2025 and 49.9% in early 2026, after the share stopped growing. Source: Graphite (an SEO agency), 2026. Limits: Graphite counted only pages marked as articles and used an AI-detection tool to sort them. Detection tools make mistakes.
9 times — Mentions of "AI slop" grew about ninefold from January 1 to November 20, 2025, compared with the same stretch of 2024. Negative sentiment peaked at 54% in October 2025. Source: Meltwater (a media monitoring company), 2025.
Mostly unseen — Graphite's own follow-up research found that most AI-written articles do not show up in Google search results. More volume does not mean more reach. Source: Graphite.
*****************************************
The volume is not the only problem. The sourcing is. When an AI system generates an answer, it pulls from whatever it was trained on — which now includes other AI-generated content. The result can be a feedback loop: unverified content produces more unverified content, each round sounding more confident than the last.
When a Wrong Answer Costs You
Most of the time an inaccurate AI answer is an inconvenience. You get the wrong museum opening date. You misattribute a quote. You walk into a restaurant that closed six months ago.
Some of the time it is not an inconvenience.
In January 2026, a Guardian investigation found that Google's AI Overviews showed "normal" ranges for liver blood tests without explaining that normal depends on age, sex, ethnicity, and other factors. Experts warned that someone with liver disease could wrongly believe their results were fine and skip follow-up care. Google removed AI Overviews for those exact searches — but similar searches still produced AI summaries for a time.
The same investigation found Google's AI advising people with pancreatic cancer to avoid high-fat foods. Experts said that is the opposite of standard medical advice. Google said its own clinicians reviewed the examples and found much of the information was not inaccurate.
This is not a hypothetical. These summaries sit at the top of Google Search, answering questions from your customers, your employees, and you.
The Due Diligence Desk is not saying AI is dangerous. The desk is saying confidence without sourcing is dangerous — and AI has made that combination available at unlimited scale.
What This Means If You Own a Business
The AI marketing hype cycle is in full swing. Agencies are rebranding overnight. The same automated blast tools are now "AI-powered outreach platforms." The same keyword reports are now "AI visibility audits." The same junior account manager is now running your "AI strategy."
Watch how the rankings get made, too. Some "top AI marketing agency" lists are published by agencies that rank themselves first. When the list maker is on the list, the list is an ad.
The questions that cut through the hype have not changed: Who funded the research behind this claim? What is the methodology? When was it published? Any vendor who cannot answer all three is selling confidence, not results.
There is a second problem specific to AI search. When someone asks ChatGPT, Perplexity, or Google AI who the best HVAC company, plumber, or roofer in your city is, the answer comes from somewhere. It pulls from reviews, directories, and other sources it treats as credible.
A business with thin, generic, or outdated information online risks one of two outcomes: it gets described incorrectly, or it does not get mentioned at all. Either way, the business loses.
The fix is not a trick. It is specific, accurate, sourced information about your business, kept current.
What This Means If You Are a Consumer
You are already living inside this problem. Every search result, product review, health answer, and local business recommendation now comes from an internet flooded with content nobody checked.
The clearest example: in February 2026, BBC tech reporter Thomas Germain spent 20 minutes writing a fake article on his personal website claiming he was one of the best hot-dog-eating tech journalists, based on a championship that does not exist. Within 24 hours, Google's AI Overviews, Google's Gemini, and ChatGPT were repeating his made-up claims as fact. The AI did not invent the claim. It found it and presented it with authority. The source was a test. The presentation looked real.
Four things to watch for:
No source named. Any answer that does not cite a specific, checkable source is a generated answer, not a verified one. "Studies show" and "experts say" are not sources.
Ranges without context. A test range, a price range, a timeframe — presented without saying what determines where you fall in it — is not useful information. It is the appearance of information.
Confidence without methodology. The most dangerous AI answers are not the ones that sound uncertain. They are the ones that sound certain when they should not be.
One source, one answer. On anything that matters — health, legal, financial, safety — check at least one primary source before acting. A named physician. A licensed attorney. A government database. The AI answer is a starting point. It is not a conclusion.
The internet has always had a spam problem. What is new is that the spam now sounds like it was written by someone who knows what they are talking about. The bar for skepticism has to rise to match.
The Due Diligence Desk's Position
The Local Aim is a local media company covering Orange County businesses. This desk applies the same standard to every claim we publish and every vendor claim we evaluate.
The Three-Question Standard
Who funded it?
What was the methodology and sample size?
When was it published?
Any source that passes all three runs as verified.
Any source that is credible but fails one question runs with this label: Directional Signal — credible but not fully verified. Test with your own data before acting.
Any source that fails two or three questions does not run.
We are not anti-AI. The tools are real. The capabilities are real. The productivity gains are real. What is not real is the claim that AI output is the same as verified fact — and much of the marketing industry treats those two things as interchangeable.
They are not interchangeable. The difference has consequences — for the consumer who acts on a wrong medical answer, and for the business owner who pays for results that cannot be measured, verified, or owned.
Vague gets people hurt. Specific gets you the call. That is not a marketing line. It is the operating principle of every verified claim on this desk.
The wave is not coming. It is already here. The businesses and consumers who know the difference between what is sourced and what is generated can make decisions from a position of clarity.
Everyone else is navigating on confidence alone.
The Local Aim · thelocalaim.com · Orange County, CA
press@thelocalaim.com · 949-832-7575Two notes before you publish: