Companies typically have cyber insurance to help them cover the costs of a data breach, but it’s worth looking to see if the policy will cover problems resulting from your use of generative AI, McCarter & English Partner Steven Weisman says.
“If AI is manipulating data and that data happens to be protected information, that could create liability risk,” Weisman told Legal Dive. “I would look for coverage under a cyber risk liability policy.
“That’s an example where, okay, you bought a cyber liability policy, but what does it actually cover?” he continued. “Did you purchase the liability side coverage? What is the scope of that coverage? Are there any exclusions that speak directly to generative AI that might cause a challenge if there’s a liability claim that arises from the handling of the data?”
A cyber policy isn’t the only insurance that might cover problems that arise from a company’s use of generative AI, Weisman said. A general liability policy might provide coverage, too.
“If an employee in the company uses information that is copyrighted or otherwise allegedly infringes on the owner’s rights to that data — an intellectual property rights or plagiarism-type claim, for example — there might be coverage for such a claim under a general liability policy,” he said.
That’s because a general liability policy typically provides coverage for personal or advertising injuries, which could include infringement of personal or business intellectual property.
Copyright infringement has become a concern among in-house counsel as they and others, as part of their work duties, step up the use of generative AI tools like ChatGPT to write first drafts of memos, court filings, contracts, training manuals and other material.
OpenAI, the developer of ChatGPT, and Microsoft, which has partnered with it on generative AI tools, have been named in at least two lawsuits by writers and other content creators seeking class action status for their claims that the companies scraped their work off the internet to train the large language models the software relies on to do its work.
News reports show concerns over copyright infringement are migrating beyond the generative AI companies to the end-users of their products, who are concerned they’re exposing themselves to liability if the court filing or press release or other material they create contains copyrighted or other protected material, like sensitive personally identifiable information.
“There will be armies of legal trolls trying to find folks to sue,” David Gewirtz of ZDNet has said.
The concern has become great enough that Microsoft recently said it would cover legal costs for users of its Copilot generative AI tools if they’re hit with infringement claims.
“We will defend the customer and pay the amount of any adverse judgments or settlements,” Chief Legal Officer Hossein Nowbar said in a Sept. 7 joint statement with Microsoft President Brad Smith.
Depending on what your general liability, cyber or other insurance policies cover, and what exclusions they specify, you might be covered for infringement by one of your insurance providers, too.
“It’s usually a puzzle,” Weisman said. “Gather the entire insurance program together and assess where coverage might exist for such a claim and take a look at each one of those policies to assess whether such a claim might be covered under one of several policies.”