Common Sense Health: These are questions for doctors, technology developers, lawmakers, educators and the public – not just for the person clicking the “I agree” button.
Children learn the rudiments of consent long before anyone uses the word. We teach them to ask before taking another child's toy, to distinguish common from personal spaces, and to understand that not everyone wants a hug or a kiss. People have boundaries, and we must respect them. As adults, we learn the principle is considerably more complicated. And these days, it’s getting hard to follow.
Sexual consent is now understood as an ongoing, voluntary agreement rather than something implied by a previous encounter, a relationship or even a marriage. In health care situations, patients are expected to participate in decisions about their care instead of simply deferring to professional authority. And in our digital lives, we are asked to consent to the collection, storage and use of personal information, often without much idea of what we are agreeing to.
Informed consent has long been central to medical ethics, grounded in the idea that people have a right to make decisions about their own bodies. Yet the shift away from medical paternalism has been significant. A patient should understand the proposed treatment, its risks, the alternatives, and the consequences of refusing it. Signing a form is no substitute for a meaningful conversation.
Now consider what happens when artificial intelligence enters the consulting room. AI can assist with interpreting medical images, identifying patterns in patient records, producing clinical notes and informing treatment recommendations. Patients may not know when AI is involved, what role it plays in a decision, or how much weight a clinician gives its conclusions.
In a September 2026 editorial in The American Journal of Bioethics, Harvard Medical School researchers Marina Hovhannisyan and J. Wesley Boyd raise concerns. Traditional consent assumes that a patient is authorizing a reasonably defined intervention. AI systems, however, can be updated, retrained and used in new ways after that initial agreement. The authors argue that meaningful consent in AI-assisted care needs to be an ongoing process, with greater transparency about the technology's role and significant changes in its use.
Consent given today cannot apply to a technology that changes tomorrow. Nor should patients be expected to understand the technical workings of a complex algorithm before they can make a reasonable decision about their care.
Health institutions have a responsibility to explain what matters in terms ordinary people can understand.
Think of the routine click on “I agree” that allows us to use an online service. We may be authorizing the collection of location data, the analysis of our behaviour or the processing of photographs and messages. How long the information will be retained? Who else might gain access? How might it be used in the future? Refusing permission invariably means losing access to a needed service. Is that meaningful consent, or the price of admission?
There is also a question of power. A patient may feel unable to refuse a recommended procedure. An employee may be reluctant to challenge an employer's request for personal information. A consumer may have little practical choice if every available service demands the same permissions. A formal yes does not establish that a person understood the implications or felt free to refuse.
None of this means that consent has become an obsolete idea. Quite the opposite. As technology gives institutions and companies greater capacity to collect information, influence decisions and act on our behalf, the principle of personal autonomy becomes more important.
But the responsibility for protecting it cannot rest solely with individuals.
These are questions for doctors, technology developers, lawmakers, educators and the public – not just for the person clicking the “I agree” button.
***
This column offers opinions on health and wellness, not personal medical advice. Visit www.docgiff.com to learn more. For comments, diana@docgiff.com.
Email Us | Reader Friendly | Comments
