Insurers are taking a cautious wait-and-see approach to artificial intelligence risks as banks and corporate clients experiment with the technology but stop short of large-scale deployment, according to senior executives at Gallagher.
The emerging consensus across the directors’ and officers’ (D&O) liability market is that AI poses a potentially transformative risk exposure, but one that remains too poorly understood for insurers to aggressively rewrite policy language or impose sweeping underwriting restrictions.
Instead, Gallagher executives said insurers are watching closely as clients test AI tools internally.
“There’s a huge amount of discussion among banks about AI implementation — who’s doing it, how they’re doing it,” said Eileen Yuen (pictured on the left), managing director - national practice leader, financial institutions at Gallagher. “There almost seems to be a race to make sure you’re involved. But discussions vary widely.”
Banks, in particular, are proceeding cautiously with AI, according to Yuen. Larger financial institutions are experimenting extensively with AI internally, while smaller and regional lenders are still determining how the technology fits into their operating models.
“The large banks, the major household names, will probably be the first to embrace it, just as they were with digital currencies and other innovations,” Yuen said. “They’ve likely spent more time internally figuring out how the metrics work compared to regional and community banks. Then the rest of the market will follow to remain competitive.
“But there’s a reluctance to fully take the leap because there’s uncertainty around both insurance coverage and regulation. Nobody really knows what the outcomes could be.”
The hesitation becomes more pronounced when AI moves closer to customers. FI clients are increasingly focused on the liability implications of using AI in functions such as mortgage underwriting, customer interactions and decision-making tools.
“That’s where I think we begin to see greater concern in the marketplace,” noted Yuen. “That could accelerate the push toward exclusions because losses will inevitably follow, as they do with any emerging exposure. Carriers are going to say, ‘Our policies can’t remain silent forever,’ because when a policy is silent, people push the language until legal precedent determines whether something is covered or not.”
The comments reflect a broader mood across the sector, with Yuen seeing parallels to the early evolution of cyber insurance more than a decade ago. “We’ve been telling clients that the most likely path forward for insurance is that, at some point, we’ll begin to see AI exclusions added to policies, much like what happened with cyber coverage,” she said.
For now, however, insurers have stopped short of imposing broad AI exclusions on financial institution or D&O policies. Policy wording, in many cases, remains effectively silent on AI-related liabilities, leaving uncertainty over how future claims may ultimately be interpreted.
The concern instead centres on what underwriters refer to as “AI washing” – companies overstating the potential impact of artificial intelligence in public disclosures or investor communications.
Jennifer Sharkey (pictured on the right), global management liability practice leader at Gallagher, said underwriters were closely monitoring how companies describe AI’s role in their business models, particularly amid soaring valuations for large technology groups.
“There have been some notable claims involving companies… with these massive AI-driven valuations,” she noted. “We’ve seen claims related to AI-washing and inflated expectations around AI, but it’s not overwhelming the market.”
Sharkey also said insurers had considered whether AI underwriting questionnaires, similar to those introduced for cyber or Y2K exposures in earlier decades, might become necessary. So far, however, carriers have largely resisted taking that step.
“We thought there would be a little bit more scrutiny,” Sharkey said. “What underwriters are looking for is just that companies are not being overly aggressive with their AI disclosures.”
That restraint reflects a D&O market that remains highly competitive even as conditions begin to stabilize after several years of falling prices.
Executives described a market still awash with capacity, particularly in higher-excess layers, though insurers are increasingly focused on profitability after years of aggressive competition eroded margins.
Public company D&O pricing in the first quarter was broadly flat, according to Gallagher data, with rates rising by roughly one percent. Market participants said that represented a significant shift from the steep declines seen over the previous three years.
“There’s still an awful lot of capacity,” Sharkey said, noting ongoing consolidation among insurers in both the US and London markets.
Insurers have continued to deploy large amounts of capital at relatively modest premiums, particularly on large corporate towers, where competition remains intense. The result has been a gradual shift towards greater underwriting discipline rather than an abrupt hardening cycle.
Some insurers have also become more selective about volatile or fast-growing companies, particularly where stock prices or business models appear highly exposed to emerging technologies.
New IPO activity is helping absorb some excess market capacity, Sharkey noted. Public listings had picked up across sectors, including aerospace, technology, retail and biotech, creating new underwriting opportunities after years of relatively stagnant public company formation.
Yet even as the D&O market steadies, brokers acknowledge that AI remains the largest unresolved question facing corporate liability markets.
For the Gallagher specialists, it’s only a matter of time before exclusions and more specialized products emerge. But until claims trends become clearer and companies move from experimentation to large-scale deployment, insurers appear unwilling to move too aggressively.
