New 'Healthy Technology Act' Lets AI Prescribe Drugs: FDA Approval and State Authorization Required

The "Healthy Technology Act of 2025" greenlights artificial intelligence (AI) and machine learning (ML) to prescribe medications, but it's not a free-for-all. This bill amends the Federal Food, Drug, and Cosmetic Act, essentially putting AI on par with human practitioners if they meet specific conditions. The AI must be both authorized by state law to prescribe and get the thumbs-up from the FDA (through sections 510(k), 513, 515, or 564 of the act). (SEC. 2.)

AI at the Pharmacy: What Could Go Wrong?

This bill's core change is massive: it legally equates AI with human doctors for prescription purposes. While the FDA approval process is meant to be a safety net, the speed at which AI is developing raises some serious questions. Can regulators really keep up with the latest algorithms, especially when a faulty one could have life-or-death consequences? For example, imagine a construction worker whose back pain gets "treated" by an AI that's been subtly programmed to favor a specific (and potentially more expensive or addictive) painkiller. Or think about a small business owner who is denied an effective medication due to biases hidden deep in an AI’s code. These aren't just hypothetical scenarios; they're real possibilities with this kind of technology. (SEC. 2.)

Patchwork Prescriptions: The State Law Problem

The bill hinges on state authorization for AI to prescribe. This means we could end up with 50 different sets of rules for how AI can dole out medications. One state might have rigorous testing and oversight, while another might rubber-stamp any AI that claims to work. This inconsistency could create huge disparities in patient care and safety, and it makes you wonder if a national standard would be a smarter move. Consider the implications for those who travel or live near state borders. (SEC. 2.(A))

The Missing Human

Perhaps the most significant concern is the removal of a human doctor from the prescribing process. While AI might be efficient, it lacks the critical thinking, ethical judgment, and ability to understand a patient's full context that a human doctor provides. Are we really ready to hand over such sensitive decisions entirely to machines, without a guaranteed human check in the system? The bill doesn't offer any insight into this.