In this bylined piece for Business Law Today, the American Bar Association's flagship business law publication, Darrow's Michael Harary argues that a new era of ERISA accountability is underway — and that the combination of expanding legal precedent and AI-powered detection is making it increasingly difficult for plan sponsors to avoid scrutiny.
Harary opens with the numbers. ERISA class action filings have grown dramatically, from just 15 cases in 2020 to more than 180 in 2025. The top ten publicly reported ERISA settlements totaled $580.5 million in 2023 — an all-time high — and $413.3 million the following year. Courts are establishing plaintiff-friendly precedents earlier in the litigation process, and the scope of claims is broadening beyond excessive fees and ESOP violations to include forfeiture practices, pharmacy benefit manager pricing, and fiduciary negligence. What was once a niche area of plaintiff-side litigation is becoming a major front.
The article then explains why ERISA violations have historically been so difficult to detect — and why that is changing. Harary identifies five structural features of retirement plans that have insulated plan managers from accountability. First, the data is complex and opaque: the information needed to assess plan performance is buried in Form 5500 filings, schedules, and attachments that lack standardized naming and require significant reconciliation. Second, investments are hidden inside complex vehicles — collective trusts, pooled accounts, and other structures that obscure the underlying assets and make it difficult to identify what a plan is actually invested in and at what cost. Additional structural barriers compound these challenges, collectively creating an environment where violations could persist undetected for years, even in the face of substantial harm to plan participants.
AI is now breaking down these barriers. By aggregating and harmonizing fragmented data across hundreds of thousands of plan filings, and applying consistent benchmarks to identify persistent underperformance relative to peer groups, legal intelligence tools can surface patterns that no individual attorney or team could realistically detect through manual review. What previously required months of analysis can now happen systematically and at scale.
Harary is clear about the implications for both plaintiff attorneys and plan fiduciaries. For the plaintiffs' bar, the combination of expanding legal theory and improved detection creates a larger universe of actionable cases than has ever existed before — including in midsized plans that were previously too difficult and costly to analyze. For plan sponsors and their counsel, the practical consequence is that the margin for error is narrowing. Decisions that might once have gone unnoticed — suboptimal fund selection, excessive fees paid to service providers, inadequate monitoring — are now more likely to be identified, benchmarked against comparable plans, and challenged in litigation.
The article closes with a straightforward call to action for the legal community: the tools now exist to protect plan participants more effectively than ever before, and the question is whether attorneys and fiduciaries will use them proactively or wait to be caught off guard.
