AI-Powered Law Firms Test the Bounds of Legal Work by Robots

Rapidly evolving AI is giving rise to a new kind of law firm: one where artificial intelligence is not just a cost-saving tool but is integral to how the firm operates.

Such AI-powered law firms are pushing the boundaries of what kind of technology is acceptable in legal work while raising hard questions about who can provide legal advice. The challenges will only grow in importance as humans cede larger swathes of legal work to advanced AI tools, including within the nation’s largest law firms.

Last month, a hybrid AI law firm called Crosby, backed by Sequoia Capital and Bain Capital Ventures, was announced officially. “We combine the speed and intelligence of AI with the safety of lawyers-in-the-loop to review contracts in under an hour,” Ryan Daniels, the company’s co-founder and CEO, said on the company’s website. Another startup, Covenant, which advertises itself as an AI law firm for private markets, said last week that it had raised $4 million in seed funding.

The upstarts are moving cautiously, for the moment. “We have to be very thoughtful about never letting AI do anything that looks like legal advice,” Daniels said. “And the case law here is still kind of opaque and still getting figured out.”

Sign up for Evan Ochsner’s weekly newsletter on legal ops & tech.

That’s because the definition of the practice of law varies by jurisdiction, and how AI use affects that is a developing issue. The American Bar Association’s guidelines on the unauthorized practice of law say “limiting the practice of law to members of the bar protects the public.”

The modernization of the legal industry and the emergence of websites like LegalZoom, which provides access to attorneys and legal services, and Rocket Lawyer, which provides online legal help, has been putting pressure on those definitions for decades.

AI is making the issue more urgent by flooding the market with advanced capabilities. An online legal service called DoNotPay that advertised itself as the “first robot lawyer” faced multiple lawuits in 2023 alleging that it was providing legal services without a license.

For the AI law firms, the patchwork of US legal regulations raises plenty of thorny questions. There’s no Supreme Court decision weighing whether a robot lawyer can be sued for malpractice, for example.

Still, legal experts say AI doesn’t take away from the basic responsibilities of lawyering.

“You have responsibility that any document that’s official that goes out with your name on it, that goes to the court or that goes to your client for use as a contract, for example, is right,” said Ted Claypoole, a lawyer who leads Womble Bond Dickinson’s IP transactions and fintech teams.

Playing It Safe

Daniels, a Stanford law graduate who worked at Cooley LLP for a year as an associate, said he is obsessed with the details of how to work with AI and yet retain the human touch.

Crosby says it is playing it safe until rules around AI lawyering become clearer.

Since January, Crosby has reviewed more than a thousand contracts like master service agreements and non-disclosure agreements, mostly from start-ups.

“We’re not trying to win necessarily on being like a super-cheap solution,” Daniels said. “I think being able to offer extraordinary speed with really good business context is really valuable.”

AI’s use by law firms could improve the average person’s access to the legal system, said Brian Liu, a co-founder of LegalZoom. How to regulate that while protecting people from bad legal advice remains a difficult question, Liu said.

“I think in general more access is better, but the public does need to be protected,” he said.

Robert J. Couture, a senior research fellow at Harvard Law’s Center on the Legal Profession, said law firms have to adopt and experiment with AI tools right now. “It’s still so early. We don’t know how it’s going to develop, but I do think that if they do not participate right now and put significant resources on this, they will be harmed,” he said.

And as for what is an AI powered law firm and what isn’t, it really doesn’t matter, he said.

“There’s no accepted definition of what artificial intelligence is today. Everybody’s got their own viewpoint of that, and that’s okay, because it’s so young,” Couture said.

Across the Pond

In England, such definitions appear to be better established, showing how regulation varies by jurisdiction. The Solicitors Regulation Authority in May cleared Garfield AI to operate in England and Wales, describing itas the first “AI-driven law firm” it had approved.

The SRA will allow Garfield AI, which helps small businesses recover unpaid debts, to provide more and more of its services without any attorney review, even though it is technically a law firm, its founder, Philip Young, said in an interview. In the two months it’s been regulated, Garfield has expanded its offerings, from pre-litigation steps like correspondence between businesses and debtors, to producing summonses and complaints, he said.

“When you create something that’s new, and no one’s ever done it before, it’s sort of incumbent on you to be prudent and responsible,” he said.

The SRA declined to comment, pointing to a statement it made when Garfield was approved. “We have been making sure there are appropriate checks in place to make sure that consumer protections are not diluted,” it said at the time.

Claypoole said advancing AI technology could force some changes in the US.

“Employers should be preparing like they did for the Internet, which is for a tool that changes things completely,” he said. “Using a hammer and nail was a way that people did carpentry for a long time, but the fact that a nail gun came along and you can use that, doesn’t change what you need to do.”