FDA's Agency-Wide AI Strategy Is in Flux After Its Biggest Champion Steps Down

The FDA's rapid AI buildout is now at risk of stalling. Former Commissioner Dr. Marty Makary, who stepped down recently, was the driving force behind the agency's centralized AI approach. His departure has introduced real uncertainty about who will carry it forward and at what pace.

What Was Built

Under Makary, the FDA deployed Elsa, an agency-wide large language model designed to accelerate drug review processes by handling support tasks like writing and summarizing reports. The agency also announced plans to expand agentic AI into premarket reviews, inspections, and administrative tasks. Between 2024 and 2025, the number of reported AI use cases at the FDA grew by more than 1,800%.

What Experts Fear

Tala Fakhouri, chief AI and regulatory strategy officer at Parexel and a former FDA AI policy official, warns that without a centralized champion, the agency may revert to a fragmented, department-by-department approach. She noted that recent FDA conference appearances have featured officials from individual divisions rather than a central AI leadership voice.

Critically, these concerns apply to the FDA's internal AI use only. Policies governing how drug companies use AI in their own work have not changed.

The Transparency Gap

Fakhouri also flagged a transparency issue: industry sponsors preparing submissions do not have clear visibility into how the FDA's AI tools analyze their packages. Greater transparency would let sponsors format data in ways that better align with what FDA reviewers and their AI tools actually need.

Key Takeaways

• FDA AI use cases grew 1,800% under Commissioner Makary; his departure creates a leadership vacuum
• Experts fear a return to fragmented, division-level AI, reversing centralization gains
• FDA policies on how drug companies use AI are unchanged

Read the full article on PharmaVoice