Biopharma misreading FDA AI guidance, former regulator warns

A former FDA official who helped write the agency's artificial intelligence policies now says the pharmaceutical industry is fundamentally misunderstanding that guidance — and creating unnecessary obstacles in the process.

Tala Fakhouri, who left the FDA last summer after working on AI policy, is now chief AI and regulatory strategy officer at Parexel, a contract research organization. From her new vantage point in industry, she's observing how companies implement the very policies she helped create.

What she's seeing troubles her, according to details first reported by STAT.

The conservative interpretation problem

Fakhouri told STAT that the FDA's intention to provide flexible guidance is being lost as companies interpret it. Rather than embracing the flexibility the agency intended, pharmaceutical companies are adopting the most conservative possible readings to minimize regulatory risk.

This overly cautious approach, Fakhouri argues, defeats the purpose of the guidance. The FDA designed its AI policies to allow innovation while maintaining safety standards, but companies appear to be prioritizing risk avoidance over the intended regulatory flexibility.

Fakhouri worked at the FDA until last summer, giving her direct experience with both the policy-writing process and the agency's goals. She shared her observations at an event in San Diego two years after her time writing AI policy at the agency.

Blame on both sides

Fakhouri doesn't place responsibility solely on industry. She indicated that both pharmaceutical companies and the FDA itself share blame for creating hurdles that shouldn't exist.

The gap between regulatory intent and industry implementation represents a common challenge in emerging technology oversight. When agencies issue guidance on novel technologies like AI, companies often struggle to determine how much flexibility they actually have versus how much caution they should exercise.

Why it matters

This disconnect between FDA intent and industry practice could slow AI adoption in drug development at a critical moment. If companies interpret guidance too conservatively, they may forgo beneficial AI applications that regulators would actually approve — limiting innovation without improving safety. The pharmaceutical industry's risk-averse culture, combined with unclear regulatory signals, can create a chilling effect that delays patient access to AI-enabled medical advances.

Fakhouri's perspective is particularly valuable because she understands both sides of the equation. Her transition from regulator to industry strategist gives her unique insight into how policy translates — or fails to translate — into practice.

These details were first reported by Brittany Trang for STAT.