Who Pays When the Robot Lawyer Errs? Crosby Buys Malpractice Coverage for Its A.I. Agents

For generations, malpractice insurance has been one of the quiet facts of legal practice. Lawyers make mistakes, clients sue, and insurers pay. Crosby, a law firm built around artificial intelligence, is now testing whether that arrangement can extend to its software. According to a report from The New York Times' DealBook, the firm has bought malpractice coverage for its A.I. agents.

The purchase is small in dollar terms but large in what it signals. It points to an emerging gap in the insurance market. Businesses are rapidly deploying autonomous A.I. tools, while the major insurers that underwrite professional risk are still working out how to handle machines that can draft contracts, flag clauses and recommend courses of action with little human involvement.

A Policy Written for Humans

Traditional professional liability policies assume a licensed person sits at the center of every decision. That person can be trained, disciplined and held to a standard of care. An A.I. agent does not fit that model easily. When an algorithm misreads an indemnification clause or misses a filing deadline, it is unclear who is at fault. The vendor, the firm and the supervising attorney could each bear some responsibility. Insurers dislike uncertainty, and the A.I. era has supplied a great deal of it.

Crosby's approach amounts to an argument that the work its agents produce should be backed the same way as work produced by associates. For clients, that offers concrete reassurance. If the software gets something wrong, there is money set aside to cover the damage, not just an apology in a terms-of-service agreement.

Insurers on Uncertain Ground

The broader industry is moving cautiously. As DealBook reported, major carriers are still grappling with A.I.-related exposure. Some are tightening policy language to exclude or limit claims tied to automated systems. Others are experimenting with specialized products. The result is uneven coverage, and early adopters are often left to negotiate their own terms.

That leaves a sector moving faster than the people who are supposed to protect it. Law firms, accounting practices and consultancies are adding A.I. agents to client-facing work. Many may not realize their existing policies were never written with such tools in mind.

Why It Matters Beyond Law

Legal services are a natural test case because mistakes there are expensive and easy to document. A missed clause or an incorrect citation leaves a paper trail, and courts have already sanctioned attorneys for filing briefs with A.I.-generated errors. If insurers can work out how to price risk in this field, the same models could spread to medicine, finance and other professions where software increasingly acts on a client's behalf.

Crosby's decision also raises a more uncomfortable question for its competitors. If one firm can show its A.I. is insured, clients may begin asking others why theirs is not. In a market built on trust, coverage may become a selling point as much as a safeguard.

For now, Crosby's policy is an early marker in largely unsettled territory. Regulators, insurers and bar associations have yet to agree on who is responsible when software does a professional's job. Until they do, firms are likely to keep patching the gaps on their own.