The American Medical Association, the largest physicians group in the United States, wants AI vendors to bear responsibility when defective clinical software causes harm, provided physicians exercised appropriate clinical judgment. Its framework allocates responsibility among developers, healthcare organizations and physicians, while protecting doctors who decline to use an AI tool or override its recommendations. The association also wants vendors to maintain insurance, disclose tool limitations and performance data, and refrain from unfairly shifting liability onto healthcare professionals through contract terms. Patients must retain avenues for recourse, while payers and hospitals should not penalize physicians who refuse to use AI they consider uncertain. This policy is not yet law: legislators would need to act for its protections to take effect.
Source: Read the original article
