Who's Liable When Your AI Gives Bad Business Advice?

You are. As of today there is no federal law that makes an AI company liable when its tool gives your business or your customers bad advice. When something goes wrong, the business that put the AI in front of people is the one who gets sued, not the company that built it.

That surprises a lot of owners. They assume the AI vendor carries some of the risk because the vendor built the thing. In practice the risk sits almost entirely on whoever deployed it, and the contract you signed to use the tool usually says so in plain language, if you read it.

Is there an actual law that makes AI liable for bad advice?

Not at the federal level, not yet. A handful of states are starting to write rules. New York introduced a bill, S7263, that would stop an AI chatbot from giving advice that would be illegal for an unlicensed person to give, and would put the liability on the company behind the chatbot if it does anyway (techlifefuture.com's liability guide). That is one state, one bill, not a settled national rule.

Outside of new bills like that one, the rules that already existed keep applying: negligence, malpractice, a business's duty of care to its own customers. Those rules were built to point at a person or a business. An algorithm cannot hold a license or carry intent, so the responsibility keeps landing where it always did, on the business running it.

So who is actually on the hook when your AI messes up?

You are, if you are the one who deployed it. If your chatbot promises a customer a refund you never approved, or your AI agent tells a client the wrong number, your business made that promise as far as the customer is concerned. The company that built the underlying model is several steps removed and almost never the one who answers for it.

This is not a hypothetical worry. General liability insurance was written before any of this existed, and standard commercial policies increasingly carry exclusions written specifically for AI related claims. When the exclusion applies, the business is left covering the claim itself, not the vendor and not the insurer.

Does your AI vendor's contract protect you?

Usually the opposite. It protects them. Most AI tools come with a terms of service page nobody reads, and that page exists to cap the vendor's own liability. When the tool gets something wrong, that contract shifts the risk onto the business running the tool, meaning you (businessinsuranceusa.com on the 2026 AI coverage gap).

There is no federal law yet that makes an AI system itself liable for the advice it gives. That responsibility defaults to whoever put it in front of your customers.

Read the liability section on any AI tool you use for anything customer facing. It almost always says the vendor is not responsible for the output. Knowing that before something breaks is a lot cheaper than finding out after.

What should you actually check before you act on AI generated advice?

1. Keep a person between the AI's output and anything that touches money, a legal commitment, or a specific client. Every agent we build for a client has a human review step before anything client facing goes out, exactly for this reason.

2. Read the liability clause in your AI vendor's terms once. Know what you are actually exposed to before you find out the hard way.

3. Write down that the review happened. If a customer ever disputes what your AI told them, having a person who checked it is the difference between a process and an excuse.

4. Do not let the AI make the final call on anything you would not let a brand new employee decide alone on their first day.

Does Your Business Insurance Actually Cover AI Mistakes? goes deeper on the insurance side of this same gap. We build the human review step into every agent we ship for exactly the reasons above; the free demo shows what that actually looks like.

---

FAQ

Can I sue an AI company if its chatbot gives bad advice?

You can try, but the contract you agreed to when you signed up almost always limits what the vendor owes you. Most of the real legal exposure lands on the business that deployed the tool, not the company that built it.

Does blaming the AI protect me if something goes wrong?

No. Saying the AI decided that does not remove your responsibility. Regulators and courts put the accountability back on the business that put the system in front of customers.

Is there insurance for this?

A few insurers launched AI liability products for small businesses in 2026, but check your existing policy first. Standard business insurance was written before AI showed up and often has gaps here.

What is the single biggest thing to fix first?

Put a person between the AI's output and anything that reaches a customer or touches money. That one step covers most of the actual risk.