| Sumario: | The application of Artificial Intelligence (AI) enabled tools is growing rapidly. These systems are being used to support complex decisions that can add value to society. In healthcare settings, AI assists in detecting, predicting, and monitoring health status, conditions, and behavior and assists in processes related to direct healthcare delivery. Amidst excitement, there is also growing concern around how AI can pose significant potential risk and harm across healthcare systems. How ought organizations, regulators, software vendors, and individual practitioners respond to these risks while still utilizing and appreciating the benefits of these technologies? This is the key ethical problem addressed by this paper. A solution is urgently needed in the AI space because practitioners and healthcare leaders are determined to realize the benefits and efficiencies of these technologies right now. Yet, standard processes that ensure human touchpoints, thoughtful consideration, and application of ethical principles to algorithm utilization are often underutilized or completely absent. This paper outlines an approach that brings communities together across healthcare IT, academia, ethics, and community advocacy to establish common foundations, articulate a framework for considering AI-enabled technologies, and train developers and purchasers of AI-enabled technologies to reflect on the ethical dimensions of these capabilities more thoroughly. We propose that a process and standards to guide the application of ethical considerations will offer guardrails in the use of this technology, heighten beneficence and nonmaleficence, introduce much-needed respect for autonomy, and enable paths for justice, especially among patients and families who will be most impacted by the inaccurate, inconclusive, or biased results.
|
|---|