<\/p>\n
Artificial intelligence is part of the fabric of healthcare. More than 80% of physicians use AI with patients, and many hospitals and health insurance companies also use it.<\/p>\n
Universal Images Group<\/p>\n
\u201cIs it OK if I turn on the AI tool to take notes? It will help me pay better attention to our conversation,\u201d her doctor said. The patient agreed; the doctor clicked a box, and the visit began.<\/p>\n
This patient is a colleague of mine, and when she recounted the situation to me, she said she felt pressured to say yes. She didn\u2019t want to slow the visit down or disappoint her doctor. And, it turns out, she\u2019s not alone.<\/p>\n
More than 80% of physicians<\/a> use some sort of artificial intelligence tool in their professional work with patients. So do many hospitals and health insurance companies, which means AI is now woven into the fabric of the healthcare system. But it comes at a cost.<\/p>\n Article continues below this ad<\/p>\n A recent Blue Cross Blue Shield<\/a> report found that hospitals and health insurers are battling it out with their opposing uses of AI. Programs that can increase health system coding and reimbursement are met by sophisticated systems to find issues with claims or expedite their review, potentially denying care. These technologies contributed nearly $1 billion in increased healthcare spending from 2023 to 2025.<\/p>\n As a physician, I understand that whether or not patients have agreed to it, they\u2019re already paying for the artificial intelligence tools in doctors\u2019 offices and hospitals. They\u2019re also too often powerless to decide what part AI takes in their care and reap little direct benefit from it. This power differential must shift.<\/p>\n Patients want to know how AI systems collect information, how their health data is being used and stored<\/a>\u00a0and how these tools impact their care. Unfortunately, the way that doctors and hospitals share this information with patients varies widely<\/a> and is often short on the details that patients need to make fully informed decisions about consent. Perhaps this is why two-thirds of adults have low trust<\/a> in the use of AI by our healthcare system.<\/p>\n Hospitals and operating rooms across the country have already turned on ambient listening<\/a> to ease the burden on nurses<\/a>, doctors and surgeons. This helps healthcare teams capture notes, diagnosis codes and other information in the electronic health record. And hospitals get paid more<\/a> when more complex and detailed information is submitted to insurance companies.<\/p>\n Article continues below this ad<\/p>\n This also may mean that when someone walks into the hospital, they must consent to this technology being always-on if they want care. Imagine feeling sick enough to go to an emergency room<\/a> only to be asked upon entry to decide if artificial intelligence can listen in, or if you\u2019d like to \u201copt-out.\u201d<\/p>\n This may be particularly challenging during a moment of crisis, though my colleague\u2019s experience speaks to how difficult it can be even during a scheduled outpatient clinic visit.<\/p>\n It\u2019s not just in the care delivery system. Eighty-four percent<\/a> of health insurance companies use AI tools to support their operations, including claims payment functions, utilization review or prior authorizations or to interface with their members. Until recently, the cost of this technology has been difficult to track, as it doesn\u2019t neatly fit into the usual healthcare\u00a0payment models<\/a> or reporting parameters.<\/p>\n Despite its cost and ubiquitous presence, there is little evidence to show that patients benefit from AI. While clinicians, who use AI tools for most of their visits, may save about 15 minutes<\/a> each day on documentation, this doesn\u2019t neatly translate to more time with patients.<\/p>\n On the other hand, there is some evidence to suggest that AI tools can\u00a0reduce clinician burnout<\/a>, which is critical to maintaining a stretched-thin healthcare workforce. Clinicians may feel they are better able to engage with<\/a> and listen to patients during visits thanks to AI in the exam room. This has an important potential impact on workforce satisfaction and sustainability.<\/p>\n Article continues below this ad<\/p>\n But when a patient seeks care from a doctor, needs to get an insurance claim paid or walks into an emergency room feeling ill, there is a power differential. If patients are to benefit from AI, they must be empowered to proactively use it to their advantage.<\/p>\n State policymakers should require health centers and health plans to rebalance their investments in AI with patients at the center, helping them to navigate the complexity of the U.S. healthcare and insurance systems.\u00a0<\/p>\n The National Academy of Medicine\u2019s framework for AI<\/a> to improve health literacy and empower patients is a critical starting point. While health systems and insurance companies have used AI to bolster their own finances, patients can become their own advocates by using AI for bill review<\/a> or filing appeals and grievances to ensure their care is paid for.<\/p>\n Guest opinions in Open Forum and Insight are produced by writers with expertise, personal experience or original insights on a subject of interest to our readers. Their views do not necessarily reflect the opinion of The Chronicle editorial board, which is committed to providing a diversity of ideas to our readership.<\/p>\n Read more about our transparency and ethics policies<\/a><\/p>\n