What You Might Have Overlooked About GPT-5

Sure! Here’s a rewritten version of your content with the promotional call-to-action adjusted accordingly:

—

With the launch of GPT-5, OpenAI has begun explicitly encouraging users to seek health advice from its models. At the launch event, Altman welcomed Felipe Millon, an OpenAI employee, and his wife, Carolina Millon, who had recently been diagnosed with multiple forms of cancer. Carolina shared her experience of asking ChatGPT for assistance with her diagnoses, mentioning that she uploaded her biopsy results to ChatGPT to help interpret medical terminology and guide her in decisions regarding treatments like radiation. The trio highlighted this as an empowering example of bridging the knowledge gap between doctors and patients.

However, this shift in focus raises concerns about accountability. OpenAI is using evidence that suggests AI can benefit doctors as a clinical tool, as seen in studies, to imply that individuals without medical training should seek advice from the AI about their own health. The danger lies in that many users might turn to this advice without consulting a doctor, particularly as the chatbot has become less prompt in encouraging such consultations.

Just days before the launch of GPT-5, the Annals of Internal Medicine published a case involving a man who ceased consuming salt and began taking dangerous amounts of bromide after conversing with ChatGPT. He suffered from bromide poisoning, a condition that had largely been eliminated in the U.S. after the FDA restricted bromide’s use in over-the-counter medications in the 1970s, and he nearly lost his life, spending weeks in the hospital.

So what does this mean? Ultimately, it’s about accountability. As AI companies transition from providing general information to offering humanlike assistance in specialized fields like healthcare, a critical question arises about the consequences of potential errors. Currently, there is little indication that tech companies will be held liable for any harm caused.

“When doctors provide harmful medical advice due to errors or biases, there’s a system in place for malpractice claims,” explains Damien Williams, assistant professor of data science and philosophy at the University of North Carolina Charlotte. “But when ChatGPT offers harmful medical advice because it’s been trained on biased data, or due to inherent ‘hallucinations’ in its operations, what recourse do users have?”

This discussion is part of our ongoing exploration into the implications of AI in society. Stay tuned for more insights from Hotnchill.

—

Feel free to let me know if you need any further adjustments!