AI Fuels ERISA Litigation as Risks Loom for Lawyers, Fiduciaries

Attorneys working on employee benefits cases are increasingly turning to artificial intelligence to crunch data, increase efficiency, and identify potential lawsuits.

Plaintiff’s lawyers are using AI tools to help identify plans covered by the Employee Retirement Income Security Act to identify outliers and anomalies and target employers with lawsuits over their handling of investment choices and forfeiture funds, while defense attorneys are turning to AI for assistance with benchmarking and comparing fees and performance across plans.

The nascent technology isn’t without risks, with benefits attorneys pointing to hallucinations and the potential for lawsuits down the road if AI is used to second-guess fiduciaries.

The steady introduction of AI tools into ERISA litigation comes as the number of class action suits has spiked so far this year, with plaintiffs testing new legal theories in lawsuits over health and retirement plans.

“You have these fragmented pieces that no one understands that’s stuck somewhere, and now we’re finally able to systematically surface that,” said Michael Harary, director of litigation partnerships at Darrow, an AI research lab focused on the legal system.

Benefits attorneys can use AI to review plan documents and identify potential violations or patterns that might lead to lawsuits. AI tools can be particularly useful for reviewing Form 5500s, the documents all large ERISA-covered plans are required to submit, benefits attorneys said.

Darrow analyzes more than 200,000 plan sponsors, 60,000 retirement funds, and more than $6 trillion in plan assets, Harary said. AI can be used to analyze the plans and each fund in the plan, then apply an appropriate benchmark against that fund, he said.

Plaintiff, Defense Uses

Many firms are already incorporating AI tools into their work.

“The emerging consensus among litigators on the plaintiff side in this space is that AI has a role in our practice,” said Adam Garner, a plaintiff’s attorney and founding member of The Garner Firm. “How we’re choosing to use it I think varies a bit based on the size of the firm, your IT budget, the types of cases you work on, and your comfort level and understanding of it.”

Garner said AI can be useful in complex cases with large volumes of documents to provide summaries or identify key information. AI has been helpful in preparing for depositions as a supplemental tool to “expand your thinking on questions that you might not have thought to ask,” he said.

Rick Nowak, co-chair of Mayer Brown’s ERISA litigation practice, said AI is being used for investment challenges and forfeiture-related cases where plaintiff’s attorneys are looking for specific markers to identify plans.

Attorneys on the plaintiff and defense sides are also relying on AI tools to help draft court filings, Nowak said, though there have been a mounting number of incidents involving AI hallucinations in filings that produce inaccurate information or case citations.

“We’re not fully there, but I think you’re just going to continue to see it grow in the litigation space,” Nowak said. “And hopefully that ultimately reduces the cost of litigation.”

Stephen Rosenberg, head of the Wagner Law Group’s ERISA litigation practice, said plaintiff’s attorneys could also use AI to generate copycat complaints by identifying recent class action complaints and sorting out ones that made it past the motion to dismiss stage.

On the defense side, Rosenberg said the technology can be used to put together a benchmarking defense for a motion to dismiss by identifying fee points and performance points of different options that a client could have selected for an investment menu “without having to laboriously pour through financial data.”

AI can also be used to spot litigation trends and determine which kinds of ERISA cases are gaining steam or have heightened potential risk for employers. Harel Fisher, who directs Darrow’s ERISA Research Lab, said data and early court rulings have shown ERISA-covered health plans are likely to be the next frontier for litigation.

Potential Risks

There are still pitfalls to using AI as the technology continues to evolve, attorneys said.

AI-produced documents and research can contain hallucinations that include misleading data, incorrect case citations, or inaccurate references to legal authority. Hallucinations in court filings have become more common, prompting some judges to fine lawyers who submit AI-aided documents containing errors.

Garner said over-reliance on AI and failing to verify the accuracy of AI tools “is a huge risk.” Attorneys should also be mindful of local rules and orders from courts regarding the use of AI, he said.

“Just like any other tool we use in our practice, you have to understand what these tools are capable of, what their limitations are, and how to use them,” Garner said.

Nowak raised lawsuits based on hindsight as a risk for fiduciaries who oversee plan assets as AI usage becomes more prevalent. Fiduciaries should be focused on process under ERISA and acting in the best interest of participants, rather than on the eventual outcome of an investment, he said.

He pointed to the Department of Labor’s recent proposal that would establish a legal safe harbor for fiduciaries who offer alternative assets in 401(k) plans and its posture in court filings as evidence the agency is seeking to provide more clarity to those overseeing plan assets in the face of rising litigation.

“The purpose of ERISA is not to second guess them six years later and say you should have done something different. AI, though, will make it easier to second-guess and rely on hindsight because it can quickly pull together things where it says, ‘Well, had you considered X, you would’ve done Y,’” Nowak said. “But that’s easy to say after the fact.”