Materiality and practicality: a response to - are clinicians ethically obligated to disclose their use of medical machine learning systems to patients?

Abstract

In his recent paper Hatherley discusses four reasons given to support mandatory disclosure of the use of machine learning technologies in healthcare, and provides counters to each of these reasons. While I agree with Hatherley's conclusion that such disclosures should not be mandatory (at least not in an upfront fashion), I raise some problems with his counters to the materiality argument. Finally, I raise another potential problem that exists in a democratic society: that even if Hatherley's (and other authors who share his conclusions) arguments are sound, in a democratic society the simple fact that most people might wish for such disclosures to be made might be an enough compelling reason to make such disclosures mandatory.