As businesses across the economy adopt artificial intelligence tools, they face heightened risks and exposures associated with the powerful technology.

One growing concern is professional liability in fields such as law and medicine, where practitioners increasingly rely on AI to enhance efficiency and performance.

Many professionals may be creating “a real exposure for errors and omissions” if they are using AI to create efficiencies without properly vetting the output to ensure it is without hallucinations or information gaps, said Maria Long, New York-based chief underwriting officer at cyber insurer Resilience.

AI can significantly heighten professional liability risk for many types of firms, said Melissa Sowa, New York-based managing director of errors and omissions product line leadership at Markel.

Concerns about AI-related professional misconduct have already reached the legal sector, said Laura Zaroski, Chicago-based managing director of Arthur J. Gallagher & Co.’s law firms practice.

“It’s not emerging, it’s definitely emerged,” in the law firm sector she said, with regular reports of “hallucinated” case law not reviewed by attorneys being filed with courts and sometimes drawing modest monetary sanctions of $1,000 or $5,000. Such fines are not covered by insurance, Ms. Zaroski said.

Although claims remain limited so far, that could change, she said. “We haven’t seen too many legal malpractice claims … I think they’re coming.”

Such claims would likely fall under legal professional liability coverage, but a rise in claims activity could prompt insurers to limit their exposure, Ms. Zaroski said.

“They’re covered like any other act, error, or omission that you would see covered under a legal malpractice policy; it’s just like having a bad first-year associate or misreading a case,” Ms. Zaroski said. “If we keep seeing an increase in claims, we may see more carriers potentially putting in some exclusions.”

Unless a policy specifically excludes AI, professionals remain responsible for mistakes caused by the tools they use, and insurers generally remain responsible for coverage, said Michael Carr, Chicago-based head of cyber portfolio underwriting, North America, for Coalition.

Coalition is paying closer attention to underwriting technology companies that offer AI products or services. “We are asking a lot more of them than we are on our typical tech errors and omissions account,” Mr. Carr said.

Standard E&O policies, including Markel’s, generally cover AI-related negligence arising from professional services, Ms. Sowa said. She added that Markel’s miscellaneous E&O form allows brokers to add affirmative AI coverage via an endorsement, subject to a sublimit.

Most E&O and cyber liability policies are written broadly enough to cover errors in professional services or security and privacy failures, regardless of the technology involved, unless they include a specific AI exclusion, said Kara Higginbotham, New York-based head of professional liability and cyber at Zurich North America.

She added, however, “there could be additional risk presented by AI that is not contemplated if there is a lack of awareness or understanding of how an insured organization is using or adopting AI when making an underwriting decision.”

As AI becomes more embedded in everyday workflows, identifying the true source of an error may become more difficult, Ms. Long said.

“It’s going to be very difficult to try to parse out how the error or omission actually happened,” Ms. Long said. “There’s a whole lot of shadow IT being used when it comes to leveraging artificial intelligence, and you can’t always tell when it’s been a part of the professional service journey, so to speak.”

Professionals must also choose the right AI tool for each task, said Michelle Chia, New York-based chief underwriting officer for cyber in the Americas at Axa XL. She noted that some AI tools are built for specific professions, including law, and that practitioners must use the appropriate application.

Failing to properly vet an AI tool and its output could be treated much like relying on untrained associates or publishing unreviewed work and could ultimately be viewed as an error or omission, Ms. Chia said.

“There are general tools that have not been trained on specific data, and so to use that tool for professional service, any type of professional service would be a disservice to the client,” Ms. Chia said.

Insurers cautious on E&O exclusions

Competitive pressures are influencing insurers’ decisions to make restrictive changes to policy wording, sources say.

Underwriters may be reluctant to add exclusions for artificial intelligence in errors and omissions policies due to concerns that their products will be less attractive in a competitive market, they said.

“You’re not going to see a lot of additional friction introduced into the underwriting process for any product until you start to see some claims, because there are commercial concerns around friction,” said Michael Carr, Chicago-based head of cyber portfolio underwriting, North America, for Coalition.

Saddling a broker or potential policyholder with another questionnaire can make an insurer less attractive, he added.

“In a soft market, it’s extremely challenging to ask any more questions,” said Maria Long, New York-based chief underwriting officer at cyber insurer Resilience.

Competition puts pressure on underwriters not to introduce additional hurdles, she said.

“If we still keep seeing this increase in claims, we may see more carriers potentially putting in some exclusions, limiting their exposure, or somehow either assertively covering it or assertively not covering it in creative ways in the future,” said Laura Zaroski, Chicago-based managing director, law firms practice, at Arthur J. Gallagher.

“It’s very competitive, so brokers are unlikely to accept AI exclusions when other insurers are offering coverage or silent regarding AI,” said Melissa Sowa, New York-based managing director, errors and omissions, product line leadership, for Markel.