Insider Insights for Multiemployer Plans | September 30, 2026 | 4-minute read
Featuring Lisa M. Gomez, former Assistant Secretary of Labor, Employee Benefits Security Administration (EBSA)
Key Takeaways
AI is probably already in your plan, through software your providers already run. The question is where it is and who is overseeing it.
Not all AI carries the same risk. Automation, analytical AI, and generative AI each call for different controls.
Ask every provider four questions: where AI is used, what data it touches, whether data is retained or used for training, and whether it assists or makes decisions.
Put AI expectations in contract provisions. Don't assume that your current contract covers it or that written protections aren't needed.
People stay accountable. Human review has to be meaningful, not simply a thoughtless click.
AI is probably already inside your plan, not because your board allowed it, but because it is inside software your providers already use. That was one of the central messages of Insider Insights Webinar #7, The AI-Empowered Plan: Efficiency, Safeguards, and the Human in the Loop, led by the Honorable Lisa M. Gomez, now President of LMG Collaborative Consulting Solutions.
As Gomez noted, the trustee question is no longer “Should we adopt AI?” It is “Where is it already, and who is watching it?”
How AI Enters the Plan
AI rarely arrives as a standalone purchase. More often, it appears through upgrades to software your service providers already run — recordkeeping platforms, claims-processing systems, document management tools, and email and productivity software. Every provider is a potential entry point.
And not everything labeled "AI" carries the same risk:
Automation follows predefined rules — routing calls, sending reminders, triggering notifications. A flawed rule can be executed consistently and at scale.
Analytical AI identifies patterns and anomalies — flagging unusual remittance activity, detecting duplicate claims, surfacing potential fraud. The risk lies in what it flags and what it misses.
Generative AI creates new content — drafting participant communications, summarizing plan documents, answering member questions. A response can sound authoritative while still being inaccurate.
When a provider says it uses AI, ask which type. The answer tells you what controls, oversight, and testing should follow.
Four Questions for Your Next Service Provider Review
Where, specifically, is AI being used? Ask for the functions, systems, and workflows affected, not a general assurance.
What data goes into the tool? Participant names, Social Security numbers, protected health information, claims history. The sensitivity of the data should drive the level of oversight.
Is data retained, shared, or used to train models? Many providers are still developing answers, which makes it especially important to ask.
Does the tool assist decisions or make recommendations? A tool that influences claims decisions or participant outcomes demands more governance than one drafting internal notes.
Put It in the Contract, Don’t Just Assume
Most AI governance expectations may fit into contract provisions plans already use: confidentiality, cybersecurity, privacy, and business associate agreements. Key provisions to consider:
Disclosure. Require providers to disclose material AI uses and give notice when they change.
Data use. Define what is permitted, including whether plan data may train models that support other customers.
Privacy and ownership. Clarify who owns the data, how long it is retained, and what happens when the relationship ends.
Security. Address incident-notification expectations, including incidents involving the provider's AI vendors and subcontractors.
Testing and human review. Set performance standards and meaningful human-review requirements for work that affects participants.
Expect transparency, testing, security, human review, and accountability, not perfection or source code.
Frequently Asked Questions
Who stays accountable when AI is involved?
People, not software. A useful framework:
AI in the loop: a person reviews output.
Human judgment in the lead: claims determinations, appeals, fiduciary decisions.
Accountability stays put: fiduciaries and providers on the hook for outcomes.
Watch for the "click" problem: approving hundreds of AI-generated recommendations an hour is a rubber stamp, not oversight.
Can AI replace the people who review its work?
No. AI should serve as a workforce extender, not a workforce replacer. Eliminating the people responsible for validating outputs removes a critical safeguard.
Your Monday Morning Checklist
Identify where AI is already being used.
Build safeguards into service-provider agreements.
Keep human judgment central to consequential decisions, including claims, appeals, disability matters, death benefits, hardship withdrawals, and QDROs.
Monitor performance, errors, and operational changes over time.
Document the process: questions asked, information reviewed, decisions made, and follow-up actions taken.
The Bottom Line
No new regulation specifically addressing AI is required to start thinking about AI in your benefit plan. Existing ERISA fiduciary duties already apply to your obligations to understand AI and how it is used in your plan. Trustees do not need to become data scientists or AI experts; they do need to ask informed questions and maintain a prudent, well-documented oversight process. Use AI to strengthen experienced people, not replace them or their judgment.
Related
About Insider Insights for Multiemployer Plans
Zenith American Solutions’ Insider Insights for Multiemployer Plans webinar series delivers timely perspectives on regulatory developments, fiduciary responsibilities, operational best practices, and emerging trends affecting Taft-Hartley and multiemployer plans. Each session helps trustees, fund offices, and industry professionals stay informed and make confident decisions in an evolving benefits landscape.
This article is based on a webinar discussion of ERISA fiduciary principles and is provided for informational purposes only. It does not constitute legal advice. Consult plan counsel regarding your specific plans and circumstances.