How Can AI Be Safely Used to Support Patients During Cancer Treatment?

The integration of artificial intelligence (AI) into health care is rapidly evolving, and within oncology, it presents both promising opportunities and complex challenges. Sally Werner, RN, BSN, MSHA, chief executive officer of Cancer Support Community, discussed the emerging role of AI-enhanced psychosocial support with CancerNetwork®.

In a recent survey conducted by Cancer Support Community, Werner reveals a striking finding: patient and caregiver attitudes toward AI in cancer care are evenly split—50% positive and 50% negative. The positive feedback centers on the potential for AI to streamline administrative tasks and improve clinical workflows, ultimately allowing for more valuable face-to-face time with their providers. Patients also see the benefit of AI in supporting clinical decision-making and optimizing treatment pathways.

However, a significant portion of respondents expressed deep reservations. Their primary concern is that AI could reduce their crucial in-person time and personal connection with clinicians. They also voiced a lack of trust in AI for making health care decisions, viewing it as a tool for information-gathering rather than as a source of definitive medical advice.

Werner notes that this cautious approach is a relief to many providers, who agree that AI is still in its infancy and should be used with discretion. The conversation also touched on critical ethical considerations, including the widespread concern about patient privacy and the need for established controls. To address these challenges, the Cancer Support Community has partnered with tech giants to develop safe, validated, and closed-system AI tools, ensuring that patients receive only proven data.

Transcript:

We did a robust research study on patient and caregiver perceptions of AI in their cancer care. Over 700 respondents shared with us that their perceptions of AI in their cancer care were about 50/50, meaning about 50% positive, 50% negative. I’ll start with the positive.

The positive we heard from patients and caregivers was that if this is going to improve the workflow of [their] provider clinician, so that [they] get more face time with them, then they appreciate that. If it helps them with decision-making and pathways, they support that. The negative components were that patients and caregivers were worried that AI may decrease their face time with their provider, their physician, their clinician, and they don’t want AI to replace their in-person time and conversations with those providers, and their connections with them. They also told us that they don’t trust AI to make health care decisions. While they may be utilizing AI to look up a diagnosis or a treatment, they don’t trust that information yet.

AI is too new, and that’s good news that they’re not trusting that information yet. Providers are relieved to hear that, too. We agree that AI is still being taught. It’s still learning. It is an information force. There are a lot of concerns around privacy. We hear that in AI and health care, AI and banking, and AI and donor relationships, privacy is something that, as a country and across the globe, we haven’t figured out how to address yet, and how we’re going to keep some privacy controls in place.

From the [Cancer Support Community] and AI standpoint, we are utilizing AI. We have an open-to-options, shared decision-making tool. We’ve had it in place for years. It’s a social work intervention, and we used a closed-system AI because we wanted to make sure privacy controls were in place, but also that patients and caregivers were only getting good, proven, validated data. We are using AI in some of our own programs and tools, but we’ve partnered with AWS [Amazon Web Services] to do that because we’re not a tech company. We’re working with one of the tech giants to make sure we’re doing this safely and efficiently.