We’ve already lived with AI applications for a good number of decades. These include a chess-playing computer system that was able to beat the world chess champion in 1997, widespread use of virtual assistants such as Siri and Alexa, targeting of advertisements online, autonomous vehicles such cars and drones, language translation, facial recognition, robotic surgery and many more.

There is little question that AI will rapidly challenge the way the healthcare industry works: from searching a person’s medical records for early signs of disease, to increasing the accuracy of cancer screening, to virtual medical assistants, to more rapid review and denials of insurance claims, AI’s impact on our healthcare system will come hard and fast. As a result, the patient-doctor relationship will be changed forever, not always for the better.

A recent survey by the Pew Research Center (February 22, 2023) found that 60% of Americans would be uncomfortable if their doctor relies on AI in managing their own healthcare. Seventy-five percent felt that AI technology is being adopted too fast. Many raised concerns about changes in personal relationships and about medical record security. A majority of participants expressed their belief that utilizing AI in healthcare will reduce medical mistakes and will lead to treating minority patients more fairly. They felt that AI could have a positive role in screening for skin cancer, but felt that it would have a negative impact on monitoring pain medication. Seventy-nine percent were against the use of chatbots in mental health care.

Another study recently published in PLOS Digital Health, showed that preference for a human clinician or an AI driven diagnostic treatment plan was evenly split in 2,472 broadly diverse participants. However, when their providers actively supported the use of the AI developed tool, participants were more likely to accept it. This points to the importance of physician interaction, validation and assurance of accuracy in driving the acceptance of AI supported healthcare.

As AI methodology and applications move into the healthcare sector, we’re facing a rapidly expanding set of new challenges with data privacy and security. Since AI systems need to be trained on massive datasets, the secure and de-identified exchange of such datasets between healthcare and AI development organizations presents a major challenge. Current regulations have not kept up with this rapidly changing field and need to be updated in order for AI to have a positive place in healthcare.

In future blogs I’ll address how AI has improved and will further benefit healthcare. I’ll also discuss the many pitfalls and shortcomings of AI, including our current unverified trust in AI technology.