Pew: Americans largely unsure about AI in healthcare

Because open-source AI models are not subject to HIPAA, disclosing medical information can open users up to data privacy risks.(CN) — Most Americans are wary of generative AI in healthcare, but a growing number use AI chatbots for medical information and advice.According to the Pew Research Center, nearly a third of Americans have used AI chatbots for medical diagnosis and treatment — and nearly all who do say it’s helpful.As world governments race to balance the risks and rewards of generative artificial intelligence, it has only become more commonplace in daily life. While opponents criticize its potential for criminal harm, AI is “revolutionizing healthcare delivery,” for better or for worse.New data from Pew suggests though some have turned to AI partly in place of traditional healthcare, most Americans are concerned about just how much AI is used each time they go to the doctor’s office.In one of two articles published Tuesday, Pew reports that more than 7 in 10 Americans say it’s extremely or very important that a doctor tell them whether AI is used in their healthcare.At the same time, only 22% of respondents said they confidently understand how doctors use AI in their healthcare, and nearly 50% said they don’t know whether it’s been used.More than half of the respondents said they don’t believe they have a say whether their doctors use AI, and would like to have more control.The average person became far more aware of just one type of AI with the release of ChatGPT in 2022. But in healthcare, AI refers to a far wider range of technologies than generative AI models.“We have been researching uses of machine learning in medical related practices for decades at this point,” public health lecturer Tim Curry told Courthouse News.Curry said he’s worked with RCG machines equipped with, by today’s standard, a basic AI model used to interpret patterns.“I think most of the world is still catching up with what AI really means, how it is really being used, and what the full impact is,” the Northern Arizona University instructor said in an email. “Especially outside of generative AI.”Still, more than 80% of respondents say they want to know whether their doctor uses AI to analyze scans and make diagnoses.“Patients have a fundamental right to know when AI touches their care,” said Jyoti Pathak, dean of Arizona State University’s School of Technology for Public Health. “Broad adoption of healthcare AI requires trust, and trust depends on clear disclosure.”AI in diagnostic and treatment algorithms comes with the most risk, Pathak said in an email. It requires “stringent FDA oversight, clinical validation, explicit transparency, and strict clinician review.”Administrative AI models for scheduling and billing, he said, are lower risk and could improve operational efficiency.Both he and Curry said education and consent are necessary for both.“These really are not separate things,” Curry said. “You would need appointments and testing to obtain the data that an AI model may be trained to help analyze.”Despite a general lack of trust in AI, more Americans seem to be turning to AI chatbots over medical professionals.In a survey of more than 4,000 Americans, more than a third of respondents said they have used an AI chatbot for medical information or advice.Of those respondents, most reported that they used the chatbot for free and affordable diagnoses. A smaller percentage said they used a chatbot to learn more about a diagnosis from a doctor or to better understand lab results.Despite young people showing more general distrust in AI, the age group most likely to use AI chatbots for medical advice was 18-29. Medical costs are rising at historic rates, and young adults have the lowest insurance rate of any other age group.AI chatbots may seem like a simple solution, but “many AI services require users to opt-out of having their chats used for training future AI models, meaning your personal health data is now being used to train the next AI advancement,” Curry said.Besides, chatbots at their current level of development can’t be relied upon for accurate medical advice.Because open-source tools often lack access to specific patient medical history and clinical contest, Pathak said chatbots are likely to persuasively misinform users, leading to either false reassurance or unnecessary panicCurry opined that medical practices could soon develop their own AI programs to discuss health information in a more controlled setting. Programs like that would be subject to HIPAA, unlike open-source AI models.In the meantime, Curry recommended that those who can’t immediately see a doctor refer to Mayo Clinic, Cleveland Clinic and other free websites with peer-reviewed information.Subscribe to our free newslettersOur weekly newsletter Closing Arguments offers the latest about ongoing trials, major litigation and rulings in courthouses around the U.S. and the world, while the monthly Under the Lights dishes the legal dirt from Hollywood, sports, Big Tech and the arts.Additional Reads