A health technology company has submitted a bold petition asking the U.S. Food and Drug Administration (FDA) to allow some artificial intelligence–based medical devices to be sold without the agency’s traditional premarket review, a move that could significantly change how AI tools reach clinical settings. The petition, filed in October by Harrison.ai, focuses on AI products used by radiologists to interpret medical images. Under the proposal, companies could release new versions of such tools without submitting full review applications — as long as the FDA has already cleared similar devices and the developer adheres to specific technical standards and post-market monitoring plans.
Exempting certain devices from the FDA’s premarket review process is unusual but not without legal precedent; the agency’s regulatory framework already allows companies to request exceptions, though such petitions are rarely used for complex products like clinical AI. If the FDA neither approves nor denies the request within 180 days, Harrison.ai’s proposal would automatically take effect — potentially as soon as mid-April this year. This timing has raised eyebrows because the push to ease regulation coincides with broader policy efforts to reduce barriers between AI developers and patients.
Critics — including some patient safety and regulatory experts — warn that shifting the burden of proof from premarket safety and effectiveness evaluations to post-market surveillance could expose patients to greater risk, especially if performance issues emerge only after widespread clinical use. Traditional FDA review is designed to catch safety or accuracy problems before a product enters the market, but with rapidly evolving AI systems, regulators and industry stakeholders have been debating whether existing pathways are well-suited to these technologies.
The petition arrives amid broader shifts at the FDA’s digital health leadership. Recently, the agency appointed a former executive from a Harrison.ai subsidiary to lead its digital health policy office — a move that has intensified scrutiny of how AI regulation may be shaped going forward. Observers see the combination of regulatory reform efforts and personnel changes as part of wider efforts to redefine how AI tools are reviewed and monitored in U.S. healthcare.
