Chamber Tells FDA: AI Pilots Should Be Voluntary, Risk-Scaled

The U.S. Chamber of Commerce is pressing the FDA to keep its planned artificial intelligence pilot for early-stage drug trials voluntary and to reward drug repurposing with a new pediatric-style exclusivity mechanism.

The recommendations came in detailed comments the Chamber submitted this spring on two Requests for Information issued by the U.S. Food and Drug Administration: one on deploying AI to sharpen decision-making in the earliest, riskiest stage of clinical research, and a second on making it easier to find new uses for drugs already on the market. Erin Delaney, Senior Director of Health Policy at the U.S. Chamber of Commerce, laid out the group's position.

Both approaches, the Chamber argues, should lead to the same destination: "a drug development ecosystem that is faster and more efficient without ever compromising patient safety or the empirical evidence standards patients and physicians rely on."

AI in Early-Phase Trials

Early-phase clinical research ranks among the most resource-intensive and uncertain stages of bringing a therapy to market. Sponsors must make dosing and safety decisions with limited information and limited patient access. FDA's proposed pilot program would explore how AI-enabled tools can support earlier, more informed decisions while remaining grounded in the National Institute of Standards and Technology's (NIST) AI Risk Management Framework.

The Chamber strongly supports the approach, with five conditions.

First, build on what exists. FDA should leverage current frameworks, guidance and pilots rather than create duplicative obligations, giving participants regulatory certainty and conserving agency resources.

Second, focus on the use case, not the technology. Expectations should scale with the risk and consequence of the decision an AI tool supports, concentrating scrutiny where the stakes are highest while letting lower-risk uses proceed efficiently.

Third, keep it voluntary. The pilot's lessons should inform future guidance through notice-and-comment and not harden into de facto standards applied outside the program.

Fourth, protect intellectual property. Participation should never require disclosing source code, model weights, or proprietary training data. Documentation-based transparency — model cards and validation summaries — can build trust without exposing trade secrets.

Fifth, keep people accountable. AI should augment, not replace, the scientific and clinical judgment of sponsors, investigators, and reviewers. Accountability for trial conduct and patient safety must remain with identifiable human decision-makers.

The Chamber also urged FDA to design the pilot so that emerging biotechs and smaller sponsors can meaningfully participate. The point, the group argues, is that efficiency gains from AI should help sponsors meet FDA's safety and effectiveness standards more efficiently — not lower them.

Unlocking Existing Drugs

"Some of the most promising treatments may already be sitting on pharmacy shelves," Delaney writes. Drug repurposing holds real promise for conditions with high unmet need — metabolic and neurodegenerative diseases, rare conditions and substance use disorders — often at lower cost and in less time.

But even for an approved medicine, establishing a new use requires meaningful clinical research. The Chamber identified four structural fixes.

Strong, well-calibrated IP protections top the list. The Chamber calls for a new pediatric-style exclusivity mechanism for sponsors who run qualifying trials to establish a new indication, modeled on the proven success of pediatric exclusivity under the Best Pharmaceuticals for Children Act.

Second, regulatory pathway clarity: clearer guidance on evidentiary standards, expanded pre-submission engagement, and better use of the 505(b)(2) pathway, so sponsors aren't deterred by uncertainty about the road to approval.

Third, flexible trial designs. Adaptive trials, master protocols, and real-world evidence should be applied aggressively where they can generate robust evidence more efficiently.

Fourth, interagency coordination among FDA, NIH, CMS, and the Department of Veterans Affairs to align development, approval, and coverage — while keeping FDA's approval decisions grounded solely in safety and substantial evidence of effectiveness, with no role for pricing or market-based factors.

The Common Thread

"A development system can be made dramatically more efficient while keeping the standards that earn public trust fully intact," the Chamber wrote in both responses.

The group's bottom line: whether advances are AI-driven or repurposing-driven, Americans have to be able to trust them. That means protecting the intellectual property that fuels private investment, keeping people accountable for the decisions that matter, and never trading away the evidence standards that protect patients.

How FDA weighs these comments will shape whether the two pilots become a template for faster cures — or another layer of obligations that slows the sponsors they aim to help.