Objections, Taken Seriously

A new word for patient power invites hard questions, and the good ones deserve real answers rather than slogans. Here are the strongest objections to patient sovereignty, stated at full strength, and answered without hedging. If any of them defeated the idea, it would be better to know now.

"Sovereignty sounds like the medical-freedom movement"

The objection is fair: sovereignty is a strong word, and other movements reach for words like it, some of them hostile to expertise and evidence. The answer is that sovereignty is a claim about who holds standing, not a claim about what is true. A sovereign nation runs the best intelligence service it can and takes expert advice constantly; sovereignty is not a license to believe anything. Patient sovereignty explicitly includes the discipline of verification, checking AI against primary sources, distinguishing a hypothesis from a conclusion, bringing findings back into a real clinical conversation. It is the opposite of the anti-evidence posture, because it makes the patient responsible for the quality of their own reasoning rather than exempt from it.

"Patients will hurt themselves"

Some will make mistakes, as some always have, and this is the objection clinicians feel most honestly. Two things answer it. First, the sovereignty is already here: about one in three adults use AI for health now, with or without anyone's approval, so the real choice is not whether patients act independently but whether the field supports them well or pretends it is not happening. Second, the sovereign patient is usually an easier patient to treat, not a harder one, arriving having read the chart and checked the evidence, which starts the conversation three levels deeper. The answer to occasional error is better tools and honest guidance, not a return to the interpretive monopoly that error rate is being measured against.

"Sovereignty for whom? This will widen inequity"

This is the objection that should worry us most, and it is not fully answered, it is a mandate. If sovereignty depends on tools, then unequal access to tools produces unequal sovereignty, and a movement that ignores that becomes sovereignty for the already-advantaged. The response is not to soften the idea but to make access central to it: the work of PatientsUseAI is precisely to widen who can exercise this standing, not to celebrate those who already can. Sovereignty is a right in principle for everyone and a practical reality for too few, and closing that gap is the point, not a footnote to it.

"This will damage the clinician relationship"

The fear is real and the opposite is true. Sovereignty ends the interpretive monopoly, but it does not end the relationship; sovereign nations sign treaties and hire advisors constantly. What changes is that the clinician stops being the sole gatekeeper of interpretation and becomes an expert partner to a person who arrives informed. The clinicians who find this threatening are the ones whose authority rested on the monopoly. The ones who find it liberating are the ones who wanted colleagues in their patients all along.

"The literature itself is AI-written now"

This is the newest objection and the sharpest, because it turns fresh evidence against the argument. If roughly nine in ten biomedical papers published in December 2025 show signs of AI-assisted writing, then a patient checking an AI answer against the literature is checking AI against AI, and the verification that sovereignty depends on starts to look like a closed loop.

Two things answer it. The first is that the contamination is symmetric and deference does not remove it. The clinician reads the same literature, follows guidelines assembled from the same corpus, and relies on reviews written the same way. Routing a patient's question through a professional does not clean the chain, it moves the machine's contribution out of view. If the corpus is compromised, everyone downstream inherits it, and the person who knows to look is in better shape than the person who was told not to.

The second is that this is an argument for the discipline rather than against it. Read the primary source instead of the abstract. Ask whether a claim rests on data or on citation. Notice when a results section carries more certainty than its methods can support, which is exactly where the study found AI signals in an estimated 58 percent of December 2025 papers. Those are learnable moves, and the objection makes them more urgent, not less available. The full evidence, including what the ninety percent does not prove, is in The Ninety Percent.

The standing challenge

If none of these objections defeats the idea, the burden shifts. See the manifesto and What Is Patient Sovereignty?