Rules for robots, and why medical AI breaks them.

Area of Science:

• Bioethics
• Health Law
• Artificial Intelligence Ethics

Background:

• General rules for protecting human rights against AI/ML tools are often inadequate.
• The White House Blueprint for an AI Bill of Rights is a recent example of an unsuccessful attempt.
• Ethicolegal analysis faces limitations when abstracting AI/ML tools from deployment contexts.

Purpose of the Study:

• To critique the efficacy of general rules for AI/ML human rights protection.
• To explore the limitations of abstracting AI/ML tools from specific contexts.
• To examine the tension between individual consent and duty-based privacy in AI medical software.

Main Methods:

• Critique of existing AI/ML human rights frameworks.
• Analysis of health technology as a case study.
• Examination of competing privacy protection models (individual consent vs. duty-based).

Main Results:

• General AI/ML rules fail to adequately protect human rights.
• Health technology highlights the conflict between privacy policy and distributional justice.
• Individual consent-based privacy can lead to biased AI training data and healthcare inequities.
• A duty-based approach aligns with critiques of prioritizing atomistic autonomy over justice and equity.

Conclusions:

• Context is crucial for AI/ML policy, particularly in health technology.
• Current privacy policies risk exacerbating healthcare disparities.
• Revisiting legal and ethical frameworks to incorporate justice and equity is necessary for responsible AI deployment.