Medical AI: is trust really the issue?

Area of Science:

• Medical Ethics
• Philosophy of Artificial Intelligence
• Trust and Reliability in Healthcare

Background:

• Hatherley's argument in the Journal of Medical Ethics posits that medical AI can be reliable but not trustworthy.
• The argument suggests trust generates moral obligations for the trustee, using patient-clinician trust as an example.

Purpose of the Study:

• To critically evaluate Hatherley's claims regarding the trustworthiness of medical artificial intelligence (AI).
• To challenge the assertion that medical AI cannot be trustworthy.
• To dispute the notion that trust inherently creates moral obligations for the trustee.

Main Methods:

• Philosophical analysis of interpersonal trust accounts.
• Critique of Hatherley's argument using established philosophical frameworks.
• Examination of the relationship between trust, reliability, and moral obligations in the context of medical AI.

Main Results:

• At least one philosophical account of interagent trust suggests medical AI can be considered trustworthy.
• The trustworthiness of medical AI is less critical than its reliability.
• Trust itself does not inherently generate moral obligations for the trustee.

Conclusions:

• Hatherley's argument that medical AI cannot be trustworthy is challenged.
• The focus on AI reliability is more pertinent than its trustworthiness.
• The claim that trust creates moral obligations for the trustee is refuted.