How AI Will Affect the Future of Your Healthcare

Big advances in healthcare are coming because of artificial intelligence: quicker, more accurate diagnoses, new drugs becoming available faster, personalized medicine based on your genes and lifestyle, and AI to prevent coverage denials.

From where I sit as a physician in Silicon Valley, I see AI reshaping what is medically possible. Think about how quickly people around you have started using AI to manage their well-being. In early 2024, only two percent of adults in the U.S. used AI for health information, according to a Salesforce survey. Now, just two years later, 29% of American adults say they use AI or a chatbot for health information every month, according to the Kaiser Family Foundation (KFF). AI will become a powerful tool for patients, as well as doctors, administrators, government and researchers. Presently:

• 40 million people around the world use ChatGPT every day for health information, according to a report OpenAI shared with Axios.
• The U.S. Government is incorporating AI into healthcare: A six-year Medicare pilot program, WISeR, uses AI to make prior authorization and coverage decisions for certain treatments in six states. 
• 71% of healthcare workers in the U.S. believe AI will be essential to healthcare operations within five years, according to another Salesforce survey.  

With all this progress, patients (and healthcare workers) still have many questions about AI in healthcare. A KFF Poll shows three out of four people worry about how private their medical information is (even though some say they’ve entered personal information while using a chatbot).  

I think it is important to keep an open mind. Artificial intelligence has been a part of healthcare for decades. Pioneers built the world’s first AI medical consultant to help with diagnoses in the 1970s. It’s clear AI is going to continue to be a part of healthcare.

With that, here is more on how AI will impact the future of six areas of healthcare: genomics, oncology (cancer treatment), radiology (a practice of medicine that uses imaging to diagnose and treat conditions), emergency medicine, drug discovery and approvals, and better administration.  

Genomics: AI will help personalize medicine for complex diseases  

With AI, physicians can offer patients highly personalized treatments based on genomics, tailoring them to a person’s individual genetics. Consider this: The human genome contains about 6 billion DNA base pairs. Scientists, such as those at the U.S. National Human Genome Research Institute (NHGRI), use AI and machine learning to analyze incredible amounts of data. This will become more complex as DNA research continues.

Here is an example. Obesity is a very complex disease, and genetics is a contributing factor for many people—there are between 200 and 500 different genes linked to obesity. Understanding which genes are involved in a patient’s case can unlock personalized treatment. Genomics, with machine learning, can help doctors understand an individual’s genetic predispositions, gene expression, the composition of their gut microbiome, behavior from wearable devices, and more, to give each patient a personalized prescription for medication with tailored diet and lifestyle plans, not a one-size-fits-all approach.

Scientists will learn gene editing tools faster and fast-track new treatments and drugs

Gene editing is still an emerging technology (with various ethical concerns that require public policy, as well as decisions in the scientific community about standards and best practices). Medical science already has gene-editing tools like CRISPR, which can study and change traits in living things. One of the most well-known uses of CRISPR is treating sickle cell anemia, a lifelong blood disease that can shorten a patient’s lifespan. Scientists have been able to take blood cells from a patient’s bone marrow, use CRISPR to turn off the genes that cause misshapen blood cells, and put them back in the patient so their bodies successfully produce healthy blood cells.

The U.S. Food and Drug Administration has now approved gene-editing treatments for people 12 and older with sickle cell anemia. It is an example of the kinds of powerful, lifesaving treatments that science can achieve with gene editing. Learning how to use CRISPR, though, is a long process. At Stanford University, researchers paired CRISPR with AI to create CRISPR-GPT, helping more scientists learn to use it and hastening the learning curve. Stanford scientists say the goal is to use CRISPR-GPT to fast-track drug development.

AI-powered genomics will predict your risk of disease

Clinicians will be able to use AI-powered genomics to predict your risk of diseases.  AI can already use facial scans to identify genetic disorders.

Some tools won’t even need to access your genes. An AI tool developed by researchers at the University of California, Los Angeles (UCLA) can detect undiagnosed cases of Alzheimer’s using a patient’s electronic health records. Alzheimer’s is the sixth leading cause of death in the U.S., and it impacts people of color, who are less likely to be diagnosed. AI tools can reduce racial disparities in diagnoses and treatment.

Diagnosing complex conditions will take less time

People with complex conditions can be diagnosed faster with AI assistance. Right now, it can take up to four years for someone with an autoimmune disorder to get a diagnosis. People who have rare conditions sometimes go through five years of visits, tests, and stress before they’re finally diagnosed. Artificial intelligence will be able to dramatically reduce the time to diagnosis, determine the best therapy for a patient, and monitor how the body responds to treatment better than before.

Using AI to study viruses will prevent disease and protect public health

Viruses have genomes, too, and AI can help study them to prevent diseases. For example, the NHGRI uses AI and machine learning to help predict future variants of the flu and COVID viruses, and those learnings support public health.

Oncology: AI will help diagnose and treat cancer earlier  

Genomics can also be used to help detect and treat cancer. Oncology is one of the areas I think we will see explode because of AI. Oncologists will be able to use AI models to identify genetic mutations and tailor treatment to each person, rather than relying on an algorithm designed to treat everyone. Google and the Genomics Institute at the University of California, Santa Cruz have developed an AI-powered tool that identifies mutations in a tumor’s genetic sequence—technology that can uncover what’s driving a patient’s cancer growth so oncologists can determine which treatments to use to address it. (Some of those treatments may include AI-assisted gene editing.) Treatments will become more effective, and with such personalized treatments, people can experience fewer side effects and better quality of life.

Already, image processing, advanced deep learning, and other AI tools can help detect skin cancer earlier and more accurately. This kind of technology can specifically help primary care physicians triage and streamline referrals, getting people access to care sooner.

Cancer screenings will be more accurate

AI can enhance many cancer screenings. Currently, AI can help reduce false positives and false negatives and direct high-risk patients into priority follow-up and care when it’s needed.

There are still some challenges, such as data bias, which occurs when AI produces unfair or skewed outputs because of the data sets it was trained on. In healthcare, that can happen when groups of people are underrepresented in data, or when cultural, genetic, linguistic, or other information isn’t captured. Sometimes AI will underdiagnose or even ignore patterns. Data bias is just one type of bias that can impact how AI works in healthcare. Health and AI professionals have developed a set of best practices to help overcome these challenges, including using data from diverse groups to develop and train AI, conducting regular algorithm audits, and educating patients and clinicians.

Radiology: Multimodal AI and superdiagnostics will lead to proactive medicine  

I believe AI will lead to huge advances in radiology. This is important progress—there aren’t enough radiologists in the United States or globally. This growing shortage can lead to delayed diagnoses and treatment, a bottleneck of cases, work overload and burnout among radiologists, less access to specialized radiologists in rural areas (which are often healthcare deserts), and financial strain on healthcare organizations. And as the population ages and people develop chronic diseases at earlier ages, the demand for radiology will increase.

Artificial intelligence can support radiologists and patients in multiple ways. By automating workflows, AI can help radiologists with scheduling, scanner selection, diagnostic support, and improving communication. AI can also help improve quality and safety, reduce exam times, and lower radiation exposure. It will improve access, especially in places where specialized radiologists are rare—global data networks will allow radiologists to collaborate. All of this is projected to save 15 million minutes of radiologists’ time every year.

These innovations are coming soon. In recent years, the U.S. Food and Drug Administration cleared more AI-enabled medical devices for radiology than for any other use.

Multimodal AI and superdiagnostics: Creating a model of you based on all your health data 

Imagine integrating data from sophisticated wearables with imaging, your genetics, your gut microbiome, lab work, health records, and more. This is the vision for multimodal AI: a health tool that analyzes data from multiple sources to develop a proactive, predictive tool for your health. Dr. Eric Topol, a digital medicine researcher, geneticist and cardiologist, told a group of doctors at the Radiological Society of North America 2024 conference, “Multimodal AI will allow us to create a high-resolution view of a human being, delivering individualized medicine that spans the patient’s entire life.”  

According to Topol, with that kind of integrated data, along with the power of AI to study and predict patterns, multimodal AI will allow doctors to understand a patient’s risk of diseases such as cancer and recommend preventative treatments and lifestyle changes before a condition develops.

Better triage and faster care in emergency rooms

When emergency rooms are overcrowded, treatment gets delayed, which can lead to worsening illnesses and injuries, worse outcomes, and even death. Crowded emergency rooms impact the entire hospital. Access to emergency care is a growing problem in the U.S. and around the world. As many as 700 U.S. hospitals, mostly in rural areas, are at risk of closing soon—for some, closure is imminent—because of serious financial problems, according to the Center for Healthcare Quality and Payment Reform, a nonprofit national policy center. Millions of Americans could lose access to emergency services or face challenges getting care.

Artificial intelligence in emergency rooms is still new, and it’s already improving care by supporting triage, managing resources, and reducing the time patients spend in the emergency room.

A study of over 174,000 visits across three emergency rooms showed that when AI assisted healthcare professionals with triage, the emergency room did a better job identifying patients who needed critical care. With AI-assisted triage, the wait time from entering an emergency department to initial care dropped 33 percent. AI also saved time by helping patients move beyond emergency care to the next stage of their treatment. This resulted in patients spending less time in the emergency room.

This doesn’t at all suggest AI should run the operation of an emergency room. In fact, the same study, published in the New England Journal of Medicine, showed that when humans collaborated with AI, healthcare performance was better than AI alone. Remember: AI is a tool and the care provider is the authority.

Faster drug discovery and approvals

Drug discovery, development, and approval notoriously take a long time. On average, it takes seven years from the first human study to the time the government approves the drug for the market. Add time for research before the first human studies, and the timeline can be as long as ten to fifteen years. Plus, the failure rate is high. Only five to ten percent of possible new drugs make it through all the clinical trials, which factors into the prices we pay for prescription medicine: The total cost of developing a new drug averages $2.8 billion.

AI has been involved in drug discovery and development since the 1980s. They were the first to use computational models. Now, though, it’s much more sophisticated.

Today, artificial intelligence can save time by quickly identifying viable options for new drugs, predicting drug efficacy and toxicity, identifying the best molecular structures for a drug, and accelerating pre-clinical testing. With AI, researchers can cut the time for discovery from an average of three to six years down to one to two years. Discovery alone is about a third of the cost of drug development. AI can save money by eliminating compounds less likely to succeed.

Once clinical trials start, AI can hasten trial length by 15 to 30 percent by helping recruit patients, monitoring them, and automating data collection and analysis.

Not only can AI save time, but it can increase the likelihood of success and prevent failures. Research published in the journal Pharmaceutics shows Phase 1 trials for drugs discovered with AI have success rates between 80 and 90 percent, much higher than the traditional average of 40 to 65 percent. And instead of as few as five percent of possible drugs making it through all the clinical trials, nine to 18 percent of drugs discovered and developed with AI assistance succeed.

Drugs will be personalized to fit the way you live

AI can also help with the pharmaceutical side of personalized medicine, in a much more holistic way. With AI algorithms, scientists can use a patient’s genetic makeup to predict how well the patient will respond to different medicines, reducing trial-and-error and one-size-fits-all treatments. This can improve patient outcomes and lead to a better quality of life. These models can also incorporate data from wearable devices (sleep patterns, activity, hydration, diet, and other lifestyle factors), a person’s social determinants of health, patient preferences, and environmental factors. Drugs won’t just be personalized based on genetics, but on a patient’s actual life.

Better billing, fewer claims denials and faster payments for doctors

AI advances healthcare efficiency in the front office, too. With AI, billing will be more transparent, prior authorizations will be faster, and both patients and care providers can minimize claim denials.  

Revenue cycle management and billing account for 25% of healthcare spending - to the tune of $496 billion a year. Every denied claim costs a provider $118—and as many as 65% are never resubmitted. Prior authorizations delay care—which can lead to worse patient outcomes, not to mention stress and financial burdens.

Machine learning can predict denials before submission, identify patterns of previous denials, and optimize submissions for success by checking for missing information, invalid patient data, or coding errors. According to a brief from the University of Colorado Denver, Kaiser Permanente reduced denials by 17 percent in six months with its pilot of an AI-powered denial management tool.  

Generative AI can create personalized cost estimates in real time before you receive care (cost transparency before getting care is required by the No Surprises Act).  Chatbots will be able to explain confusing billing details and offer personalized payment plans or financial assistance to patients who need them.

Healthcare workers are optimistic about AI. As of 2025, 63 percent of healthcare organizations used AI and automation to help manage revenue cycles, according to the Healthcare Financial Management Association, a nonprofit for healthcare finance professionals.

Automating billing and revenue processes could save the healthcare industry billions a year—money that can go back into improving facilities and equipment, expanding services, staffing, and better care.

In my small practice, I can see that AI has exploded, and I see no signs of it slowing down. I think it’s important for doctors to let patients know how they’re using AI. I also think it’s important for everyone to remember that artificial intelligence is a tool, and the doctor remains the authority.