Broad artificial intelligence exclusions are appearing in management liability policies, creating a potential coverage gap for employers using automated hiring tools as discrimination litigation tests who is responsible when algorithms produce biased outcomes.
The issue is gaining urgency as Mobley v. Workday moves forward in federal court. The proposed class action alleges Workday's AI-powered screening tools discriminated against applicants based on age, race and disability.
In June 2026, a federal judge rejected much of Workday's attempt to dismiss amended claims, allowing key California and federal disability allegations to proceed. Workday has denied that its tools make hiring decisions or discriminate against applicants.
As the litigation develops, the insurance market's response is fracturing three ways. Some carriers are attaching absolute AI exclusions to management liability packages. Most policies remain silent, treating algorithmic and human misconduct alike.
For now, AI exclusions remain uncommon in employment practices liability insurance. According to Lucas Roberts (pictured on the right), a management liability broker at Burns & Wilcox, they appear in closer to 10% of the EPL market. Their breadth, however, means even incidental use of AI could become central to a coverage dispute.
“We’ve been seeing absolute AI exclusions, with carriers right out of the gate carving that exposure completely out,” Roberts told Insurance Business.
“Should a claim allege at any point that AI was used to discriminate against someone, the carrier could say, ‘Our absolute AI exclusion precludes us from considering this because we’ve specifically carved that exposure out, so we’re off the claim.’”
Roberts said the wording is generally placed in the common terms and conditions of a management liability package, rather than added specifically to EPL coverage. It could affect D&O, E&O, fiduciary or employment claims connected to AI.
At the other end of the market, the majority of policies draw no line between an algorithm's decision and a hiring manager's. Chris Williams (pictured on the left), employment practices liability product manager at Travelers, said AI-related claims have emerged but are not yet numerous.
“The vast majority of EPL policies do not make a distinction between wrongful employment practices committed by humans versus AI,” he said. “Thus, absent having a specific AI exclusion on its EPL policy, if a company was sued for discriminating against an applicant or employee due to its AI guidelines, the coverage would respond the same as if the discrimination had been committed by a human.”
The policy silence is a red flag for employers because automated systems can replicate patterns in historical hiring data without being explicitly instructed to favor or reject a protected group. Roberts described how a system designed to identify candidates most likely to be hired could learn from previous selections and recommend applicants with similar characteristics. “By doing that, it could accidentally create a discriminatory practice,” he said. “The open question is whether Workday is then responsible for the discriminatory practice that the AI created, even though Workday didn’t tell the AI to do it.”
A handful of insurers are moving in the opposite direction from the exclusion writers. Some carriers such as Counterpart are now offering affirmative wording clarifying that coverage can respond regardless of whether AI contributed to the alleged conduct.
Meanwhile, regulators are imposing safeguards in some jurisdictions.
New York City requires certain automated employment decision tools to undergo bias audits, while employers must publish audit summaries and notify affected candidates.
Colorado’s revised AI law establishes disclosure, recordkeeping and human-review requirements for automated systems used in consequential decisions, including employment, beginning in 2027.
However, until the regulatory landscape becomes clearer, coverage analysis will turn on the policy language and underlying allegations. Roberts cautioned against relying on informal assurances that an AI exclusion is intended only for technology or D&O claims.
He added that courts examining a denial generally focus on the policy and the claim, not informal discussions about what an exclusion was supposed to accomplish. If the protection is not in the wording, it may not exist at all.
"If a claim arises from using that technology, you want to make sure that the carrier has your back," Roberts said.
