Organizations are deploying AI tools faster than governance frameworks can keep pace. The technology is advancing quickly, but the structures designed to oversee it are still evolving. In many cases, organizations are still working to define what effective AI governance looks like in practice.
These challenges were the focus of the American Arbitration Association® (AAA®) panel discussion, “Trust Is the Product: How AI Governance and Legal Oversight Are Driving Successful Adoption,” at Legalweek 2026. Moderated by AAA President and CEO Bridget McCormack, the panel brought together perspectives from legal, technology, and business leaders, including Galia Amram (OpenAI), Anna Gressel (Freshfields), and Henry Hagen (Moderna).
Drawing on these perspectives, the discussion explored how organizations are designing AI governance frameworks in practice, highlighting the evolving role of legal teams, the importance of cross-functional collaboration, and the emerging challenges posed by agentic AI systems.
Governance Is Evolving Fast
Legal teams are no longer just gatekeepers. They’re becoming enablers of responsible AI use, helping organizations move forward safely while keeping pace with innovation.
In practice, governance is changing in a few key ways:
-
From static frameworks → to agile, principle-based approaches
-
From one-time review → to continuous monitoring and lifecycle management
-
From siloed oversight → to cross-functional collaboration
The Next Frontier: Agentic AI and Commerce
While many organizations are still refining governance for generative AI, the next challenge is already emerging: agentic AI systems.
These systems don’t just generate outputs; they can act, make decisions, and interact with other systems autonomously.
This raises entirely new governance questions:
-
Who is accountable for an AI agent’s actions?
-
Where should human oversight sit in automated workflows?
-
What controls are needed when AI systems can execute tasks, not just recommend them?
And increasingly, this is moving beyond internal operations into agentic AI commerce, where AI agents may transact, negotiate, and make purchasing decisions on behalf of organizations.
Governance models built for human-in-the-loop systems won’t be enough. The next phase will require embedded oversight, defined authority boundaries, and new approaches to accountability at scale.
What Organizations Should Do Now
The organizations getting this right aren’t waiting for perfect frameworks. They are:
-
Starting with real business problems
-
Building governance that aligns with strategy
-
Continuously monitoring AI systems after deployment
-
Engaging legal, technical, and business teams together
Because governance isn’t about slowing innovation, it’s about enabling it.
Download the Full Report
To explore these insights in more detail, download the full report: