AI Risk & Liability
You used the tool. The output went out the door. Now someone is looking at you.
Get the AI Liability Exposure MapThe problem is not “AI risk.” It’s ownership.
Most solos and small businesses treat AI like a helpful assistant. Courts and regulators are starting to treat the output like something you said or sold.
That shift is already visible:
- Customer-facing chatbots are being called products at the pleading stage.
- State Attorneys General are using ordinary deceptive-trade-practice laws against AI companies — and the same theories work against the people who deploy the tools.
- Insurance carriers are quietly adding generative-AI exclusions.
- Vendor contracts still try to leave you holding the bag.
Florida is not waiting for a perfect AI law
Florida does not have a comprehensive AI statute. What it has is more practical — and more dangerous for the unprepared:
- FDUTPA (Florida’s deceptive and unfair trade practices act) is already the primary weapon. Private plaintiffs can recover actual damages plus attorney’s fees. The AG can seek up to $10,000 per willful violation.
- The Florida Attorney General filed the first state-led lawsuit against OpenAI in June 2026, then sought a temporary injunction in September. The theories (deceptive marketing, negligence, product liability, public nuisance) are portable.
- Deepfakes and nonconsensual altered images carry real criminal and civil consequences. Political AI content requires a specific disclaimer.
- Lawyers remain under Florida Bar Ethics Opinion 24-1 — the first AI ethics opinion in the country.
→ Read the full Florida overview (free, source-linked)
Actually Florida-specific tools
Sound familiar? The questions that keep people up
- Do I have to tell my clients I used AI?
Depends on context and profession. Political ads in Florida require a specific disclaimer. Professionals (especially lawyers under Florida Bar Ethics Opinion 24-1) have heightened duties around confidentiality, verification, and informed consent. Unsubstantiated claims about what the AI can do are the bigger risk. - How do I know what problem to fix first?
Start with the recurring bottleneck that drains energy without creating client value. Clarity before tools. - Should I fix my process before automating it?
Yes. Automating a broken process multiplies the brokenness — and the liability surface. - What should I automate first?
High-frequency, low-judgment, energy-draining tasks with clear inputs and outputs. Keep judgment-heavy and relationship-critical work under human control. - When AI gets it wrong, who pays?
In the current climate, the working answer is you — the deployer. Vendor indemnity is usually narrow. Check your policy for generative-AI exclusions before the next renewal. - Who owns what you build with AI?
Pure AI output still has no human author for copyright purposes. The contract language around training data, outputs, and derivatives is where ownership is actually decided. - Why time management isn’t solving your time problem.
Because the problem is usually priority and process design, not calendar blocks.
What actually reduces exposure
- Treat every customer- or client-facing output as if you said it yourself.
- Stop repeating vendor capability claims in your own marketing.
- Read the vendor terms for data use, training rights, and indemnity before you scale.
- Pull your insurance and look for generative-AI exclusions (CG 40 47 and equivalents).
- Keep a simple record of what tools you use for what work and who reviews the high-stakes pieces.
Start here
Want help applying it?
Sometimes you just need someone who has seen how these problems actually play out and can help you see the next clear step. No hype. No doom. Just clarity.
The Working Session — $250. One hour, one-on-one, your direction.
Book the Working Session →John Dennison is a retired Florida attorney. Nothing on this page is legal advice or creates an attorney-client relationship. Informational only. Last reviewed: October 2026.