Using ChatGPT in Your Business? Here Are the Insurance Risks You're Probably Ignoring

If you search “ChatGPT business insurance risk,” you'll mostly find articles about how AI is disrupting the insurance industry. That's not what you're looking for.

You want to know: does using ChatGPT in my business create insurance problems?

Yes. Here's how.

The Problem: Your Insurance Wasn't Built for AI

General liability (GL) policies were designed to cover traditional business risks: slip-and-fall, property damage, advertising injury, professional errors. They were never designed to cover liability created by artificial intelligence.

Until recently, that didn't matter much. AI was niche. Now, with ChatGPT, Copilot, Midjourney, and dozens of other tools, AI is embedded in daily operations at most small businesses. And insurers have responded.

In January 2026, Verisk — the organization that writes the standardized policy language used by most U.S. carriers — released three new endorsement forms that give insurers the ability to explicitly exclude AI-related liability from your GL policy:

• CG 40 47 — Total AI exclusion (nuclear option)
• CG 40 48 — AI liability limitation (caps and restrictions)
• CG 35 08 — AI definition and scope (defines what counts as AI)

AIG, WR Berkley, Great American, and Hamilton are already implementing these. More carriers are expected to follow. For a deep dive on how the exclusion language works, read our Verisk CG 40 47 explainer.

5 Real Ways ChatGPT Creates Insurance Risk

This isn't abstract. These are specific, everyday uses of ChatGPT that create liability your insurance may no longer cover.

1. Client Proposals with AI-Generated Errors

Your sales team uses ChatGPT to draft client proposals. The AI hallucinates a capability your product doesn't have, or misstates pricing from an old data set. The client signs based on that proposal. When the deliverable doesn't match, they sue for misrepresentation.

Under a standard GL policy, this might be covered as an advertising injury or professional liability claim. With a CG 40 47 endorsement on your policy, the carrier can deny coverage because the error originated from an AI tool.

2. Marketing Copy with Copyright Issues

Your marketing team uses ChatGPT to write blog posts, social media content, or ad copy. The AI reproduces copyrighted material — a passage from a published article, a tagline from a competitor, or lyrics from a song. You publish it. The rights holder sends a cease-and-desist or sues.

This is an advertising injury claim. If your carrier has adopted AI exclusion endorsements, they may argue the liability arose from AI use and decline to defend you.

3. Customer Service Chatbots Giving Bad Advice

You set up a ChatGPT-powered chatbot for customer support. It tells a customer that your product is safe for a use it's not designed for. The customer relies on that advice and suffers harm or financial loss.

This is a classic liability scenario — except now the “employee” who gave the bad advice is an AI. Under Verisk's definition (CG 35 08), chatbots are explicitly covered as AI technology. Any claim arising from the chatbot's output could be excluded.

4. Confidential Data Leaked Through AI

An employee pastes customer data, financial records, or proprietary information into ChatGPT. That data is now in OpenAI's system. If it surfaces in another user's output or is part of a data breach, your business is liable under data protection laws.

This creates both a cyber liability claim and a potential GL claim. The cyber policy may cover some of it, but if your GL carrier can point to AI involvement, they have new tools to limit or deny coverage.

5. AI-Assisted Code with Licensing Problems

Your developers use GitHub Copilot or ChatGPT to write code. The AI reproduces open-source code under a restrictive license (GPL, AGPL). You ship that code in a commercial product. The license holder files a claim.

Intellectual property claims from AI-generated code are a growing category. If your professional liability or GL policy carries an AI exclusion, you're potentially uninsured for this risk.

The New Reality: HSB Launches AI Liability Insurance

On March 18, 2026, HSB (Hartford Steam Boiler) launched a dedicated AI liability insurance product specifically for small businesses. This is significant — it confirms that the insurance industry sees AI as a distinct, insurable risk category, separate from standard GL coverage.

In other words: the market is telling you that your current insurance probably doesn't cover AI risks adequately. Dedicated products are emerging because the gap is real.

What You Should Do (Before Your Next Renewal)

You don't need to stop using ChatGPT. You need to document and govern how you use it. Here's the minimum:

The Bottom Line

ChatGPT is an incredible productivity tool. It's also creating a new category of business liability that most small businesses aren't governing and most insurance policies weren't designed to cover.

The insurance industry has already moved. Verisk's endorsements are filed. Carriers are adopting them. HSB just launched a dedicated AI liability product. The question isn't whether this affects your business — it's whether you'll be ready when your renewal hits.

The businesses that document their AI governance now will have leverage at renewal. The ones that don't will find out the hard way that “we use ChatGPT for everything” isn't a risk profile any underwriter wants to cover.