The present and future of artificial intelligence (AI) in healthcare

Dr. Alan Karthikesalingam MD PhD has led the research team for Google Health UK for nearly 4 years. Throughout his time on the team, he’s led and published a variety of research exploring artificial intelligence (AI) in healthcare. We asked him a few questions…

Google Health: Thanks for chatting with us Dr. Alan! Let’s start off with an “easy” question - what are some of the problems that exist in healthcare today that artificial intelligence can help solve in the future?

Dr. Alan: Happy to chat though that’s not an easy question to start with! There are many problems that AI might be able to help solve in the future. AI can empower billions of consumers to make more informed decisions, and can give clinicians and patients back the gift of time by building useful assistants for caregivers around the world. For example, diagnostics and screening programs around the world face a shortage of trained experts, from mammography to diabetic retinal screening, digital pathology, and much more - AI can help prioritize the most urgent cases.?

Google Health: Got it. So how is AI making the jobs of doctors a bit easier?

Dr. Alan: AI is able to assist doctors with tasks they already perform as well as assist experts with repetitive or exhausting tasks. To give just one example - planning a radiotherapy case requires hours of meticulous “contouring” of tissues. AI assistance can take the time required to complete these plans from hours down to minutes - freeing up doctors’ time for more complex and nuanced clinical decision-making.

AI can also help to assist doctors in screening and procedures - for example, acting as a navigation aid or teammate in settings like colonoscopy, or assisting in the interpretation of complex medical imaging.?

Google Health: So we know how AI is helping doctors. How is AI helping patients better manage their health?

Dr. Alan: AI can help provide authoritative information to people, so that they can make informed decisions about their health. AI can assist us to lead healthier lives, prevent the onset of disease and enable its earlier detection by augmenting both everyday consumer technology and medical screening programs alike. In just one example, the Fitbit Heart Study has shown how AI in everyday wearables can help millions of Fitbit users detect undiagnosed heart rhythm problems (like “atrial fibrillation”) with high reliability, enabling many to seek appropriate care and hopefully preventing future complications like strokes.

Google Health: At Google Health, we believe that healthcare equity is one of today’s most critical issues. How can AI make healthcare more equitable?

Dr. Alan: If developed and deployed inclusively, thoughtfully and with care, AI might be able to improve accessibility and accuracy of healthcare for underserved populations, neglected diseases and more. Significant care is needed to address systemic societal challenges, health data poverty and underrepresentation, and to employ procedural and technical safeguards to counteract the risks of bias in AI systems to ensure they do not propagate or amplify existing inequities.?

Google Health: So there’s a lot of things that AI could help with! With that, humor us. What do you think health AI will look like in 10 years?

Dr. Alan: In 10 years, I hope to see useful machine intelligence that complements caregivers and consumers alike, adding value to society. We will have discovered ways that AI can assist in prevention of illness, and improve the cost-effectiveness, clinical-effectiveness and quality of experience of treatment for those who need it.

For more information on Google Health’s work in AI visit: https://health.google/ ?