If you run a family law or estate planning firm with two to ten attorneys, you’ve seen the headlines: judges sanctioning lawyers over the misuse of AI. And somewhere in the back of your head, you may have decided that means AI isn’t for your firm.
I want to show you why that’s the wrong lesson — and I want to start with a conversation that actually happened.
"Using AI is fraud"
A marketing team at a firm we work with told me, flat out: “Using AI is fraud.”
Here’s the thing. This firm prosecutes fraud. So when they say that, they mean it.
And they’re not alone. Read the comments under any of these sanction stories and you’ll find potential clients saying things like, “If I found out my lawyer was using ChatGPT, I’d fire them so fast.”
Where they're right
About legal work? They’re right.
If an AI drafts a brief, nobody checks it, and it goes in front of a judge with made-up cases in it — that deserves sanctions. Someone hiring a lawyer to protect their kids or their freedom does not want a chatbot advising them. Full stop.
So if you’ve been hesitating on AI, that’s not you being behind. That’s your judgment working.
But here’s what the headlines never tell you about what actually happened in those cases.
This is a legal malpractice story, not an AI story
Honestly, when I see these stories, I roll my eyes a little. Not because the sanctions weren’t deserved — they were. It’s because of the lesson everyone takes from them.
Nobody in these cases got sanctioned for using AI. They got sanctioned for signing words they never checked. That’s a proofreading failure wearing an AI costume.
You’d get the same sanction for citing a book you never opened. AI is a book that talks back — you still have to check the citations. The attorneys in the headlines skipped due diligence, and the AI took the headline.
In every one of these cases, the judge’s question is the same: “Where did this come from?” The sanctioned attorneys couldn’t answer it.
I call that the provenance rule: know where every word came from.
And here’s the turn. That exact question — the one that sank the attorneys in the news — is what makes AI marketing safe, if you build your system around answering it.
The Closed Book Method
First, understand what AI actually is, because it’s what the lawyers in the headlines got backwards. AI is not a source of knowledge. It’s a presenter of your knowledge. You confirm the facts. It makes them presentable.
Here’s how we run it at IFTS. We call it the Closed Book Method.
Most people use AI like an open-book exam — except the AI brings its own book, and sometimes the book is made up. Those are the “AI hallucinations” you’ve read about, and they’re exactly where the bogus citations come from.
We run it closed book:
- We sit your attorneys down on camera — about thirty to sixty minutes each — and interview them about the questions their prospects actually ask.
- That footage becomes the only book allowed in the room. The AI uses only what the attorney said on camera — nothing from outside it — to draft the blogs, the videos, the website content, all of it.
- Then a human does the final pass. AI gets you eighty percent of the way; a person finishes the last twenty. Every piece, no exceptions.
If you didn’t say it on camera, the AI can’t say it either. It can’t cite a bogus case, because it was never allowed to go looking for cases. It can only repackage what you said.
So when the provenance question comes, the answer is always the same: From the attorney. On camera. We have the footage.
A real firm, real numbers
I know the theory only goes so far, so here’s a real example.
Bononi & Company is a Pennsylvania firm covering bankruptcy, family law, estate planning, and business law. Thirty minutes of filming per attorney fuels about six months of marketing material for their entire website. Across a year, that’s roughly eight hours of attorney time.
What came out of it:
- Over 1,500 qualified leads
- 75,000+ new website visitors
- Best campaigns at about $49 per lead
- A top blog post ranking on page one of Google with over 8,000 views
- Three new attorneys hired to handle the caseload
And here’s the whole story in two sentences: the lawyers in the headlines got sanctioned for filing words they never verified. Bononi doesn’t even have to approve their marketing content — because there isn’t a word in it they didn’t say themselves.
That’s the provenance rule, running in a closed-book environment.
Where we draw our own lines
So you know we hold ourselves to the same standard: our videos that use an AI twin get marked as AI on YouTube, every time. Our 80/20 content doesn’t — because by our own rule, the knowledge is entirely the human’s. Nobody discloses that a paralegal formatted their newsletter. This is the same thing.
Two things you can use today
Before any AI-assisted content publishes at your firm, run the two-question check:
- Does it cite anything specific? A law, a stat, a case, a deadline.
- If yes — can you find that exact fact in a primary source? Not by asking the AI to confirm itself. A primary source.
Can’t find it? That’s the failure from the headlines, caught before it cost you anything.
And second: name a sign-off owner. One person — even in a three-person firm — owns the final check before anything AI-assisted goes live.
One quick flag while you’re at it: YouTube’s AI labels and your state bar’s advertising rules are two separate tracks. Content you said on camera and approved sits comfortably inside both — but check your state bar’s rules. That’s a one-hour project, not a reason to opt out of AI.
The bottom line
The firms sitting this out because of the headlines are handing an advantage to the firms that built the checks.
The question was never “is AI safe?” It’s the judge’s question: do you know where every word came from?
If your answer is “from the attorney, on camera, and we have the footage” — you will never be the one making the news.
Want to see exactly how this works? The full Bononi & Company case study is right there — no form, no call, no email gate. Just a link, and it works.
