The Brief:
US AI law firm Crosby has insured its AI agents with a US$5m malpractice policy, believed to be the first of its kind.
Its agents now handle ~20% of the firm’s work with no lawyer sign-off.
New York AI law firm Crosby has secured a US$5m malpractice insurance policy for its AI agents. As far as the firm knows, it’s the first of its kind.
The firm already runs most of its contract review through AI, with ~60 lawyers checking the output. NDAs come back in as little as an hour, for a flat fee.
But founder and CEO Ryan Daniels says his agents are ready to work solo.
In some cases, they did better work than our lawyers.
Getting covered took six months. Crosby eventually signed with Testudo, a two-year-old insurer backed by Lloyd’s of London.
When the policy went live on 3 November, Crosby pulled lawyers off work that doesn’t legally require licensed attorney oversight, or “paralegal tasks”. That includes reformatting documents and checking them against a playbook, and makes up ~20% of its workload.
Anything that’s clearly legal work still needs a lawyer. Practising without a licence is illegal across the US, and California has just barred lawyers from delegating the practice of law to gen AI.
Daniels doesn’t think that will last.
“We assume that at some point in the next year or two, our associations are going to start to relax what agents can do, because they can clearly do legal work,” he said.
Big insurers, meanwhile, are backing away from AI risk. As of April, at least 16 of the 20 largest US liability insurers had sought permission to exclude generative AI from their policies.
And 7 of 13 major legal malpractice insurers have seen a rise in AI-related claims over the past year, including hallucination cases.
Insurers usually only back legal work with a lawyer’s name on it. Now they’re backing agents directly. On routine work, lawyer review starts to look less like a safeguard and more like a cost.
Source: The New York Times
