I get it—it’s super tempting to save yourself some time by popping a quick prompt into ChatGPT or Claude and pasting the results. Maybe you’re extra cautious and do a little fact check before you hit Ctrl-P and publish your latest blog. I hate to be the one to tell you that you’re probably posting misinformation. And AI is so good at sounding authoritative that you won’t even realize it.
Whether you’re a lawyer, a paralegal, or a legal content writer, you know that you have to meet ABA advertising requirements. When something else is writing the content for you, the little inconsistencies are easy to overlook. A little “experienced lawyer” here, some “guaranteed results” there. What’s wrong with a little marketing hyperbole? Unfortunately, even the most cautious AI user can run afoul of ABA marketing guidelines and good legal practice. Worse, AI slop and misinformation can tank your search engine rankings and harm your firm’s reputation.
If you’re thinking about using AI, including large language models (LLMs) for your website and blog content, consider the following problems before you invest your time and money.
AI Hallucinates Constantly
AI is very good at sounding authoritative, but it doesn’t think. It is also not designed to say “I don’t know”. An AI agent will literally make up untrue facts, statutory references, case law, and expert opinions just to avoid saying it doesn’t know something. Even the most cautious prompts can’t train it away.
Several lawyers have learned this the hard way. Briefs, appeals, complaints, and other court documents created by AI are rife with nonexistent laws and court cases. In 2026, you can hardly open a news page without another story about a lawyer being sanctioned, fined, or held in contempt for filing inaccurate or erroneous briefs.
Damien Charlotin, an expert in legal data analysis and the use of LLMs in the legal profession, maintains a database of legal decisions involving AI hallucinated content worldwide. The US leads with 1,395 reported cases of fabricated, misquoted, and misrepresented case law that resulted in sanctions, fines, and other penalties against the responsible attorneys.
AI Doesn’t Know Case Law
There’s a reason AI is hallucinating these cases; it doesn’t understand case law. It doesn’t understand anything, because it doesn’t actually think. LLMs predict how words relate to each other, using existing content to predict words and sentences based on statistical likelihood. AI models do not, however, understand how those words and sentences fit into a larger, cohesive narrative. They can’t make solid arguments or research supporting material. They can, however, make up authoritative-sounding word salad that looks legit at a glance.
Even with paid access to LexisNexis, Westlaw, and other legal databases, LLMs don’t know how to read through legal documents to understand how a judge arrived at a decision. They certainly don’t know how to relate the reasoning to an existing case and write a compelling, accurate argument. Which is why they make things up.
AI Steals Material
It’s essentially plagiarism software, and it’s stealing from you and your competitors already. In some cases, word-for-word. Without a human checking every sentence, you don’t know if you’re saying something someone already said. And let’s be honest—that’s exactly what you’re doing with AI anyway. The software doesn’t create anything new; it scrapes content from existing websites and then uses glorified text prediction to create something that sounds real and authoritative.
If you’re writing a legal blog or web page, you’re looking at extra liability for plagiarism. You’re also not going to get the kind of rankings and credibility you need to stand out from the competition. Especially if you’re using the competition’s words and ideas.
AI Costs You More Than Just Using People
When LLMs first came on the scene, they were billed as the affordable writing option. Just enter your prompts, they said, and we’ll give you low-cost words that rank for SEO. Trust us. Trust our LLM to do the drafting and the thinking and the analyzing. It didn’t take long for people to realize that AI was making things up. So, companies kept laying off writers while keeping “editors” to fact check.
Then, people started to notice that LLMs sounded… robotic, oddly enough. Editors were asked to make the AI material sound more human. Writers were hired back to “humanize” AI content. When writers pointed out that correcting AI took more time and effort than writing from scratch, companies looked away. It was just a growing pain, a small bump. The amount of money spent on AI programs would definitely reflect a solid return on investment, any day now.
Now that everyone has laid off their writing and editing teams, the cost to keep AI tokens flowing has exceeded the cost of just keeping those employees. That doesn’t include the amount companies are paying to hire in-house or freelance writers and editors to fact-check AI content.
The numbers reflect these issues too. A 2025 study by MIT shows that 95% percent of organizations who invested in GenAI saw zero return on $30-40 billion in enterprise investment. As of September 2026, executives attending the WSJ Technology Council Summit said that for most users, AI is “wasting more money that it’s worth”. Some leadership teams think this can be corrected with better training and upgraded tech. However, industry leaders have recently called to slow AI development, which doesn’t bode well for the upgrades users are hoping for. We won’t get into the veracity of those claims.
AI Slop Is Getting Deprioritized
Before you copy-past that AI-written blog, remember that Google’s search quality raters rank your pages based on E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness). For YMYL (Your Money or Your Life) content, Google prioritizes material offering actual value and accurate information where incorrect info can lead to real-world harm.
When your content is incorrect, poorly researched, or repetitive, Google’s raters rank it lower, which means your site will fall behind better-written pages for the same searches. Since most people only look at the first page of Google results, you lose significant visibility without a decent search ranking. With Google’s AI summaries taking up more of that click volume, authoritative, expert-written content is even more important for getting the traffic you need.
AI Is Making Itself Worse
We know that AI feeds off the internet. LLMs take existing material, analyze it, and decide which words or sentences should come next. With more and more unchecked AI material out there, the LLMs are regurgitating incorrect facts over and over until they seem like they must be real. Everyone’s saying it, right?
As a writer for legal websites and blog content, I’ve spent hours tracing facts and statistics only to discover that there is no original, valid source for the information being disseminated. When you’re publishing legal website content, you can’t afford to have misinformation attributed to your firm. At best, it compromises your reputation and credibility. At worst, AI material opens you up to ABA fines and penalties and even litigation.
If you’re considering using AI for your legal content, ask yourself: Is it really worth the risk to save a few bucks on the front end? Would you even be saving money?