How ChatGPT May Have Saved My Life
Mon Jan 12 2026
In a world where AI sits in our pockets and answers our questions in seconds, how do we know when to trust the technology with our personal health? On today’s episode, host Avis Favaro talks with Alvina Nadeem, an engineer, mother and ovarian cancer survivor who says AI likely saved her life when she input symptoms that she had ignored into ChatGPT and it flagged the possibility she had cancer.
But researchers from the University of Waterloo, including Sirisha Rambhatla, Director of the Critical Machine Learning (ML) Lab at the University of Waterloo, caution that using artificial intelligence to self-diagnose illness can misinform. They found that AI was right about a third of the time when prompted using real-world questions.
Bottom line: AI can be a powerful tool for health care, but experts urge it needs some kind of guardrails to prevent it from doing more harm than good.
The episode is available in English only.
Episode transcripts can be found on our podcast web page.
Learn more about the Canadian Institute for Health Information (CIHI).
Say hello on LinkedIn, Facebook and Instagram, or subscribe to our YouTube channel. You can also stay up to date on the latest news and upcoming releases with our monthly newsletter.
Vous préférez les balados français? Voyez le Balado d’information sur la santé au Canada.
More
In a world where AI sits in our pockets and answers our questions in seconds, how do we know when to trust the technology with our personal health? On today’s episode, host Avis Favaro talks with Alvina Nadeem, an engineer, mother and ovarian cancer survivor who says AI likely saved her life when she input symptoms that she had ignored into ChatGPT and it flagged the possibility she had cancer. But researchers from the University of Waterloo, including Sirisha Rambhatla, Director of the Critical Machine Learning (ML) Lab at the University of Waterloo, caution that using artificial intelligence to self-diagnose illness can misinform. They found that AI was right about a third of the time when prompted using real-world questions. Bottom line: AI can be a powerful tool for health care, but experts urge it needs some kind of guardrails to prevent it from doing more harm than good. The episode is available in English only. Episode transcripts can be found on our podcast web page. Learn more about the Canadian Institute for Health Information (CIHI). Say hello on LinkedIn, Facebook and Instagram, or subscribe to our YouTube channel. You can also stay up to date on the latest news and upcoming releases with our monthly newsletter. Vous préférez les balados français? Voyez le Balado d’information sur la santé au Canada.