Labor practitioners who have incorporated artificial intelligence into their practices say these developing tools can crunch data and expedite tedious tasks even as they remain conscious of its shortcomings and the perils of misuse, though other attorneys say they're using AI minimally or not at all.
In August, the NLRB rejected a fire sprinkler company's challenge to an agency judge's ruling and referred the consultant who led it for sanctions for citing to precedent and transcript passages that were made up by AI. (iStock.com/Sandwish)
While some in the practice remain skeptical, labor attorneys on both sides told Law360 they've found value in using artificial intelligence tools to analyze documents, research legal issues and expedite other common, time-intensive tasks. At the same time, these early forays have shown that lawyers must take care when they incorporate AI into their work, sources said.
Artificial intelligence "can be very helpful," said Sarah Goodman, a principal at management-side
Offit Kurman.
"It can help speed up the mechanical parts of the work, but it does not replace judgment on strategy, credibility or how a case can be argued," Goodman said.
OpenAI released ChatGPT nearly four years ago, giving many members of the public their first brush with a new class of advanced artificial intelligence tools — including chatbots, text generators and task-performing systems called agents — that approximate aspects of human reasoning and creation. Since then, the technology has spread across numerous areas of society, the legal industry included.
Morgan Lewis & Bockius LLP, one of the top management-side labor firms, has embraced AI tech.
The firm has found AI tools to be a natural fit for its labor practice, said Josh Rosenzweig, who helps lead Morgan Lewis' deployment of AI tools across its numerous practice groups as the firm's senior director of AI and innovation. This is because representing businesses in labor relations involves frequent decision-making on fast-moving events, informed by large amounts of data and recurring patterns, he said.
"Where we've really found AI to be helpful is organizing and analyzing that data alongside our experienced labor lawyers who are responsible for the advice," Rosenzweig said. "So what we have really started to do is pair those things together."
For example, Rosenzweig said, an employer's collective bargaining agreement may be years old and the managers who negotiated it may not be around when the time comes to hash out a new one. Morgan Lewis attorneys can reconstruct the history of these agreements by tasking an AI tool with analyzing prior deals and drafts, bargaining notes and internal communications. This can help negotiators understand how the expiring agreement came together, guiding the employer's pursuit of its goals in the next contract.
The firm also employs AI tools to help clients administer active contracts, including by synthesizing contract language and arbitration decisions to help interpret provisions and flagging concerning patterns for employers by tracking grievances and tying them to the relevant contract language, labor partner David Broderdorf said.
But AI has significant limitations for labor lawyers, especially its difficulty parsing the nuances of the law, Broderdorf added. Those nuances are many in a practice with a deep, 90-year body of law that can be conflicting and difficult to understand. Because of this, labor attorneys can't rely on AI to reach conclusions or provide practical advice, he said.
"Citing cases and starting out with a research framework, yes, but reaching conclusions as to what's lawful, what's not lawful, what's likely to lead to litigation, what's not likely to lead to litigation?" Broderdorf said. "It's been less effective there, and needs to be managed very closely by labor lawyers."
Practitioners on the labor side
appear generally less enthusiastic about AI tools. James Shaw, a partner at union-side
Segal Roitman, said he does not use AI tools in his practice. Nor is he aware of other labor-side practitioners who openly use AI, though he suspects that some colleagues use it but don't admit it.
Other union-side sources said they use AI minimally or not at all, but the technology does have boosters in the union movement, with attorney and policy commentator Matt Bruenig chief among them.
"[AI] has dramatically changed what I do and the number of clients that I can have at any given time," said Bruenig, a solo practioner. "I've reworked most of what I do around this."
Bruenig, a prominent left-wing voice on social media platform X, worked for unions and the
National Labor Relations Board early in his career before launching a think tank called the People's Policy Project in 2017. He continued to work on the periphery of the labor movement but is now once again a "legit working labor lawyer" representing multiple unions, he said. As part of that work, he has developed a suite of AI tools that he uses in his own practice and sells exclusively to unions and union-side lawyers.
Among these tools is a data compendium called NLRB Research made up of six databases. These include a searchable collection of the entire historical corpus of National Labor Relations Act law — including NLRB and court decisions and guidance materials such as general counsel memos — that subscribers can query with an AI tool. Bruenig has also compiled similar troves of Merit Systems Protection Board, Labor-Management Reporting and Disclosure Act, and
California Public Employment Relations Board law, as well as collective bargaining agreements and public filings to NLRB dockets.
Bruenig also publishes an AI-powered newsletter called NLRB Edge that summarizes new legal developments with occasional commentary, and consults with unions to organize their collections of CBAs, which for some unions number in the thousands.
Among the other AI tools that Bruenig uses in his practice is an agent that logs his hours and assists with invoicing, but his most extensive use of AI for substantive legal work has been to assist brief writing, he said. Bruenig still writes his own briefs but leans on AI for parts of the process, such as adding citations.
"I'm writing my brief, I know what I'm referencing is in the transcript because I was in the hearing, but I don't want to go through 800 pages," Bruenig said. "[I say,] 'Hey, read that sentence, find the transcript cite,' and then it does."
Bruenig confirms the responses manually, but still saves "a good five minutes" on every citation, which adds up when briefs may include hundreds, he said.
A recent NLRB decision appears to illustrate the dangers of failing to check what AI tools say. In August, the board rejected a fire sprinkler company's challenge to an agency judge's ruling and referred the consultant who led it for
sanctions for citing to made-up precedent and transcript passages. These errors suggest "the document … was at least partially generated by artificial intelligence," the board said.
But while this appears to be an example of AI leading a party astray, it was not an error unique to the use of this technology, Offit Kurman's Goodman said.
"I don't think the board's frustration was really about AI," Goodman said. "It was about a filer not doing the basic professional duty of confirming that the authority cited actually exists and says what it's cited for. And that's an obligation that predates AI."
Like Bruenig, Goodman said she is careful to double-check AI results. She added that practitioners and clients must be aware that they risk waiving privilege by uploading documents to public AI tools or bringing them into legal discussions. Her firm has amended its engagement letters to clients to address this concern, she said.
Aleksandr Felstiner, a partner at union-side
Levy Ratner PC, said union attorneys have an obligation to their clients to understand how employers use AI for hiring, performance evaluation, discipline and other personnel matters. As for his own use, Felstiner said his firm has been "tentatively exploring" the use of AI tools in its work but declined to share specifics.
Those explorations have shown the technology may be helpful in certain areas, such as performing legal research,
drafting routine communications and filings, and helping attorneys organize and query their own materials, he said.
Felstiner noted that AI has also bled into "every product that attorneys already use," including research platforms and systems for managing documents. Because of this, AI appears unavoidable in a labor practice. But how essential it is, and for what purposes, are still mysteries.
"I have to assume [AI tools] will become an everyday part of every attorney's practice, but I don't know exactly where they will fit in," he said.
--Editing by Bruce Goldman and Amy Rowe.
For a reprint of this article, please contact reprints@law360.com.