Centering Patients' Voices in Artificial Intelligence‒Based Telemedicine

• Published in: American journal of public health (1971) 2023-05, Vol.113 (5), p.470-471
• Main Author: Kraft, Stephanie A
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Chatbots; Correspondence; Decision making; Ethics; Fairness; Health care; Health care delivery; Health disparities; Health education; Health services; Humans; Interpersonal relations; Medical decision making; Medical technology; Medicine; Patients; Public health; Quality of care; Receiving; Technology policy; Telecommunications; Telemedicine; Transparency; Trust; Trustworthiness; Values
• ISSN: 0090-0036; 1541-0048
• DOI: 10.2105/AJPH.2023.307270

Artificial intelligence (AI) and telemedicine, individually and together, are rapidly changing how patients interact with the health care system. AI coupled with health care delivery via telemedicine can help promote shared decision-making by bringing datainformed predictions to patients and their providers.1 These technologies have potential to improve meaningful access to quality care by meeting patients where they are and empowering them, in collaboration with their providers, to take steps toward promoting their health.Yet AI-powered health care also has potential to exacerbate existing health care disparities,2 cause harm to patients, and fail to serve the health of the public if it is not designed and deployed with attention to patients' values, needs, and priorities. In doing so, these technologies may not only cause harm but also may undermine public trust in AI in medicine. The recent proliferation of AI-powered chatbots illustrates the simultaneous power of AI and its potential threats to trust in science and medicine. The journal Nature, for example, recently issued guidance for transparency around the use of chatbots in producing scientific work out of a recognition that their use could signal a lack of scientific integrity and trustworthiness.3AI in health care delivery similarly has great potential to support clinical communication and decision-making, but it is easy to imagine how its use could convey insincerity, provide inaccurate or unhelpful information, or otherwise undermine a therapeutic patientprovider relationship. Indeed, trustworthy AI in medicine requires attention not just to considerations such as transparency of the AI itself but also to how its use affects patients' experiences receiving care, their ability to make autonomous decisions, and health equity across society more broadly. Trustworthy medical AI must incorporate a nuanced understanding of people's health care experiences-which previous qualitative work has shown can be influenced by multiple factors at both the interpersonal and organizational levels4-and how AI may change them. A holistic approach that recognizes the interplay among technology, policy, and interpersonal relationships is essential to understand how AI affects patients' experiences and ability to trust in the care they are receiving.