Artificial intelligence is quickly becoming part of everyday health decisions. Millions of people now consult AI tools like ChatGPT for explanations of symptoms, medical information, and possible diagnoses. But recent studies have raised concerns that chatbots may sometimes give misleading or even dangerous medical advice.

In this episode of LSAT Logic Applied, Andrew Leahey examines the reasoning behind those warnings. Studies have found that AI systems can misidentify conditions or recommend the wrong next steps in simulated medical scenarios—but does that necessarily mean the technology is unreliable in real-world situations?

Using classic LSAT logical reasoning tools, this episode explores a key question: when researchers test AI in controlled experiments, how confidently can we generalize those results to everyday use? Along the way, we examine a common reasoning flaw involving experimental conditions, alternative explanations, and the difference between identifying a potential risk and proving a broader conclusion.

The goal isn’t to decide whether people should use AI for health advice. Instead, the focus is on how arguments about new technologies are constructed—and how careful logical analysis can reveal what the evidence actually supports.

Even if you’ve never taken the LSAT, the reasoning skills used in this episode apply to debates about technology, medicine, and scientific claims more broadly.