A prominent AI-powered law firm’s plans to buy liability insurance for its AI outputs is raising questions about evolving human oversight of legal tech.
The announcement by venture capital-backed law firm Crosby marks a departure from the tenuous line legal tech companies have walked by touting what AI can do while also driving home the notion that lawyers should continue to review the technology’s outputs. Liability insurance may help insulate firms from civil suits to a degree, but the looming issue for state bar associations, courts, and regulators is where the line should be drawn between what a chatbot can do on its own and what a lawyer has to review.
More legal tech companies are likely to start taking a more cavalier approach to lawyering without lawyers, Oliver Roberts, who teaches about AI and the legal industry at Washington University in St. Louis, said in an email.
“We should expect more legal technology companies to pursue this path in the coming months and years,” Roberts said.
At the same time, insurance for agents can’t fully insulate an AI-powered law firm from consequences, said Vincent Rezzouk, a privacy, data, and AI partner at Bird & Bird. In most jurisdictions, law firms still have to be owned by lawyers, so even as AI becomes more autonomous, there’s still a lawyer somewhere to take blame, he said.
“We can artificially pretend that it’s all agents and there’s no lawyer,” Rezzouk said. “But I think that at this stage, there is always a person behind. There’s always a responsible lawyer who will take the overall responsibility and who will be responsible and subject to discipline if things go wrong.”
‘Autonomous Legal Work’
Crosby, backed by Sequoia Capital and Bain Capital Ventures, will take out professional liability insurance for its AI agents so they can do “autonomous legal work,” founder Ryan Daniels said in a LinkedIn post this week.
A short video accompanying the post suggests that a user could ask Crosby a question about a contract and quickly get an “insured output” from one of Crosby’s agents.
“Today, our lawyers review every single work output,” Daniels wrote in the post. “As agents have improved in leaps over the last few months, it’s become clear this won’t be necessary in the future.”
Lawyer review of AI outputs remains the overwhelming standard throughout the legal industry. New law firms building around the technology still employ teams of lawyers to review outputs, while legacy firms require their lawyers to verify filings as they adopt more AI tools.
Daniels also described a far more cautious approach to Bloomberg Law last year.
“We have to be very thoughtful about never letting AI do anything that looks like legal advice,” Daniels previously said. “And the case law here is still kind of opaque and still getting figured out.”
In an email, Daniels declined an interview request and said the company would have more to share in the coming months.
Jake Sendar, who founded a tech company that automates some of the legal work tied to transactions, said liability insurance would represent a departure from the perceived appeal of AI-powered law firms: that a human touch is part of their core offering.
“I’ve always understood the main appeal of AI-native services to be that they provide the best of both worlds: the efficiency and cost savings of AI combined with the judgment and accountability of experienced professionals,” Sendar said in a post.
Crosby is suggesting that lawyers might not be central to lawyering after all, he said.
“They might be right, and in many ways this is the natural end-state of a truly AI-native law firm,” Sendar said.
Unprepared Regulators
Liability insurance doesn’t address the issue of whether those agents can, by law, provide legal advice.
“The problem with the insurance question is that it totally begs the question: What are you insuring against?” Anthony Davis, a legal ethics lawyer and partner at FisherBroyles, said in an email. “Are you insuring against a malpractice action? If it’s a malpractice action, it assumes that the AI provider is practicing law.”
States generally restrict non-lawyers from practicing law and providing legal advice. However, there isn’t a clear definition of what constitutes the practice of law. AI has only brought further confusion to an issue that has never gotten a clear answer, whether its been raised by authors who have written books that look like legal advice or by online services like LegalZoom and Rocket Lawyer.
Crosby will work with auditors, bar associations and state regulators, Daniels said in his announcement.
Some efforts to clarify the robot lawyer question have emerged. New York lawmakers last year introduced a bill that would impose liability for damages if a person used a chatbot in a way that would impersonate licensed professionals like lawyers. OpenAI has also argued in a lawsuit that users aren’t supposed to use ChatGPT as a lawyer.
As it stands, regulators and bar associations have focused more on regulating lawyers themselves than the thornier and broader questions about legal services, Davis said.
“This finally presents everybody with the problem that the current system of regulation doesn’t fit what’s happening,” Davis said.