| Sumario: | This article critically examines the ethical framework for AI-enabled clinical trial recruitment proposed by Rentzepis et al., highlighting a key structural vulnerability related to the lack of enforceable transparency for commercial AI recruitment tools. It emphasizes that while Rentzepis et al. identify transparency challenges and recommend academic reporting standards, they do not specify mechanisms to ensure commercial vendors disclose algorithmic details, which are often protected as trade secrets. The authors propose a concrete solution: conditional regulatory approval requiring post-market, third-party audits of recruitment algorithms to assess model architecture, demographic performance, stability, and agreement with human eligibility determinations. This approach aims to uphold ethical principles of respect, beneficence, nonmaleficence, and justice by enabling verification of AI tools used in regulated trials, addressing bias, and ensuring equitable participant selection.
|
|---|