- What is GPT-4?
- But does GPT-4 actually know anything about medicine?
- An AI for medical experts and non-experts alike
- A new partnership with AI raises new questions
- Back to Zak and his mother
This chapter is from the book
This chapter is from the book
This chapter is from the book
A new partnership with AI raises new questions
By now I hope you are getting the idea that GPT-4 is a whole new kind of software tool. Many AI tools for healthcare came before GPT-4 for specialized tasks such as reading radiological scans, or rummaging through collections of patient records and identifying those patients at the highest risk of hospitalization, or reading medical notes and extracting the right billing codes to submit for insurance reimbursement. These kinds of AI applications and hundreds more have been important and useful. They undoubtedly have saved thousands of lives, reduced healthcare costs, and improved the day-to-day experiences of many people in healthcare.
But GPT-4 is a truly different breed of AI. It is not a system specially trained for any specific healthcare task. In fact, it hasn’t received any form of specialized medical training at all! Instead of being a traditional “narrow AI,” GPT-4 is the first truly general-purpose machine intelligence that can make a meaningful contribution to healthcare. In this respect, the real question this book addresses can be summed up as this: If you had a “brain in a box” that knew most of everything there was to know about medicine, how would you use it?
However, there is another even more fundamental question. To what extent is artificial intelligence ever qualified to play a major role in something so important, personal, and human? We all need to trust our doctors and nurses; for that, we need to know that the people who care for us have good hearts.
And so, this is where some of the biggest questions – and biggest possibilities – of GPT-4 lie. In what sense is GPT-4 “good?” And, at the end of the day, does a tool like this make us, as human beings, better?