A startling revelation has emerged from a recent survey conducted by the Canadian Medical Association, shedding light on the growing reliance of Canadians on artificial intelligence for health information. The survey, which surveyed 5,000 Canadians, reveals that approximately half of the population is turning to AI for health-related queries, a trend that has raised concerns among medical professionals. This shift towards AI for health information is particularly concerning given the significant disparity in health outcomes. Those who trust AI for health information are five times more likely to report adverse health effects compared to those who do not. The survey also highlights the widespread use of the internet for health information, with nearly all Canadians searching online and 64% encountering false or misleading content. Interestingly, the primary reason for seeking health information online is the perceived speed of access, but a significant 57% of respondents admitted to using the internet only when unable to access a family doctor or other healthcare professional. Despite the caution and skepticism surrounding AI, the survey indicates that Canadians are increasingly using AI platforms like ChatGPT and Google AI summaries. However, a staggering 73% of respondents do not trust AI to provide accurate health information, yet they continue to use it. This paradoxical behavior raises questions about the trustworthiness of AI in the healthcare domain. The survey further underscores the erosion of trust in health information sources, with 77% of Canadians expressing concern about health misinformation from the United States, leading to increased skepticism even towards reputable sources. When it comes to trust in specific sources, Canadians are less confident in news organizations and provincial public health agencies, with 40% displaying neutrality, skepticism, or outright distrust towards scientific studies. In response to the growing reliance on AI, the Canadian Medical Association emphasizes the need for better access to healthcare for all Canadians. The association's president, a medical oncologist, highlights the anxiety caused by patients using AI for diagnoses and treatment advice, citing a study where ChatGPT only correctly diagnosed half of the medical cases presented to it. This study, published in the journal PLOS One, underscores the potential pitfalls of AI in medical diagnosis. The survey's findings also highlight the increasing trend of patients seeking health information from family members, while still maintaining trust in healthcare practitioners as experts. However, the potential risks of AI's expanding influence in the healthcare system cannot be ignored, as evidenced by the Guardian's investigation into Google's AI Overviews, which found inaccurate and potentially harmful information. The survey's implications extend beyond individual health decisions, emphasizing the need for a comprehensive approach to address the challenges posed by the integration of AI in healthcare.
