AI Empowers Patients to Navigate Cancer Care

A panel of experts discussed the promise and challenges of using artificial intelligence (AI) in cancer research and care during a March 24, 2026, virtual forum for patient advocates presented by the American Association for Cancer Research (AACR).

The discussion, moderated by Anna D. Barker, PhD, FAACR, chief strategy officer at the Ellison Medical Institute, brought three cancer researchers and a patient advocate together to discuss how AI will help guide decisions for patients, researchers and physicians.

Using AI to Navigate Cancer Treatment

Many patients are already using AI chatbots, such as ChatGPT, Claude, Gemini, and Perplexity, to summarize complex medical information from their patient records, noted stage 4 triple-negative breast cancer survivor Saranne Rothberg. Rothberg, who is the founder and chief executive officer of the Comedy Cures Foundation, said that AI chatbots allow users to quickly break down diagnosis and treatment information.

In addition, chatbots can function as AI agents, tools designed to solve problems and perform tasks rather than just provide information. For instance, a patient might use a chatbot as an agent to match their specific medical profile to available clinical trials. That patient can follow up with further prompts to refine information to their needs.

“I asked my AI agent to find me every trial that was appropriate for my cancer,” Rothberg said. “That agent is working to decipher that [information], and then I can just ask it again.”

Rothberg said that patients are using AI agents to help them write appeals for insurance denials, helping them grasp the scientific concepts of a case and make more compelling arguments. But some insurance companies are using AI as well to generate claims denials, said George W. Sledge Jr., MD, a breast medical oncologist at Stanford Medicine Cancer Center and chief medical officer of Caris Life Sciences. “Five years from now, what we’re going to have is dueling artificial intelligence deciding [insurance claims].”

Preparing for Appointments Using AI

Sledge remembers a time when physicians acted as a “mystic priesthood” that owned all medical data and knowledge. “That era disappeared some time ago, largely because of the internet,” he said. “But we’ve now approached a point where patients come into my clinic having asked any of the AI chatbots what is the most appropriate therapy for their disease.” This technology also enables patients to research complex conditions and therapies, mirroring how internet search has helped patients find treatment information for years.

“They come in with a fairly sophisticated analysis of what they should be treated with,” Sledge said. “Now, those analyses are not always correct because the prompts are not always perfect, but the patients are certainly much more educated at every level than they were before.”

Carlo C. Maley, PhD, a biologist specializing in cancer, evolution, and computational biology at Arizona State University, noted that patients who are preparing to meet with their doctors can ask AI chatbots to prepare specific questions for the appointment.

“AI is very good at tutoring and supporting [patients],” Maley said. “I see it not just sort of on the research side but empowering that [doctor-patient] relationship.”

Oncologists will also learn to rely on AI to empower their decision-making, said Alan Russell, PhD, head of research and development at Amgen.

“Oncologists do amazing, incredible work, but an oncologist with the power of AI is likely to do even better work,” said Russell. “A patient with AI is going to be able to interface and collectively find that wisdom that lets them make the decision with their oncologist. I think AI upskills both sides of the equation.”

Ensuring Human Oversight in AI

While panelists said they see enormous potential for AI in health care treatment and research, they also noted that the tools come with some risks. Rothberg said patients should remove personal information like names, next of kin, and social security numbers from queries.

In addition, AI sometimes hallucinates, producing convincing but false information. Sledge cautioned that as AI approaches human-level intelligence, it may become even more adept at lying.

At the beginning of the discussion, Barker asked, “In the future, should AI become more successful at predicting, detecting and curing cancer than oncologists can, will human medical judgement become irrelevant?” The panelists gave a unanimous “No.” While AI may eventually excel at choosing initial therapies, the experts agreed that human judgment remains essential to manage biases.

“These are practical ethics that come up 10 to 20 times a day if you’re in a doctor’s office, and to blithely say that artificial intelligence is going to be 100% predictive ignores the realities of what happens in every clinic every day,” Sledge said.

Overcoming the AI Digital Divide

Barker likened the adoption of AI to GPS navigations system in cars—once a novelty, but now an everyday part of driving to unknown locations. Rothberg said she asked her health care team if AI had assessed her scans as well as the radiologist.

But some populations may have no idea what AI entails or how to use it. In fact, some people may just have started using search engines like Google to look up more information about their cancer.

“There’s going to be a socioeconomic divide, and then there’s going to be an age divide for older citizens who are like, ‘I’m not going to get involved in that AI stuff,’” Rothberg said. “I think we have to start thinking about the discrepancy, how fast it is evolving and how many people are getting left behind.”

This article was originally published on the Cancer Today website. Cancer Today is a quarterly magazine for cancer patients, survivors, and caregivers published by the American Association for Cancer Research. Subscriptions to the magazine are free.