The AI Revolution in Medicine: GPT-4 and Beyond

Gave it to a daughter who initially was very sceptical. Changed mind through 180 degrees!

The power of AI promises to optimise healthcare and medical research around the world which gives anyone hope looking for cures to be developed for untreatable diseases more quickly and efficiently. It doesn't shy away from the dangers of AI and assuming it doesn't wipe us all out it gives a clear glimpse of how it could make healthcare so much better for patients and increasingly stressed medical staff. I found this book well written and easy to read and recommend it to anyone interested in this fascinating subject. It feels like we are living in the Stone Age compared to what is around the corner!

I found this book to be highly informative and well researched, with clear examples of the impact of AI on medicine, particularly focused on Chat GBT. The authors are all top notch--one is the leading research on AI in healthcare from Microsoft, one is a leading bioinformatics expert and physician from Harvard Medical School and the other is one of the best science and healthcare journalists in the country. Very refreshing to see a book that keeps the hype factor to a minimum and delivers the goods on actually explaining why this all matters to patients, physicians, and those running healthcare provider organizations. The book provides an unvarnished and pragmatic view of the high impact of AI bots and models on medicine and healthcare, with a particular focus on what this will mean for patients, doctors and the healthcare provider community. The view of AI from the patient's point of view is refreshing, since much of what is written about healthcare technology is usually so much mumbo jumbo about standards, protocols and other IT systems plumbing issues. This book will be very useful to doctors who will need to understand how to use AI as a valued partner and to patients who will need to understand how to use AI to manage their own health and medical literacy. The authors are very clear about the limitations of the existing AI systems and encourage everyone to "trust but verify" which is very good advice. I found the detailed chats in the book to be focused enough to follow and clearly illustrate the main points. Carey Goldberg, the journalist who worked on the book, does a very good job of cutting through the jargon and presenting the information from experts Peter Lee and Zak Kohane in a very clear, easy-to-read and understand fashion. Wish there was a book like this for many other aspects of AI chatbots for use for writing, research, everyday life, etc. Well worth the investment of time and attention to understand how AI will impact medicine and healthcare, which is important for patients, physicians, and policy makers alike. Hope this will be an ongoing series of books as generative AI and the use of artificial intelligences evolved and become increasingly important in our daily lives.

Maybe one day also in german translation? AI / (Chat-) GPT and beyond will help doctors and patients to get to a new level of medical information - far beyond Google adwords! Everybody should read...

the recent advance of open AI is incredible. It seems to be able to understand human emotion. I want to share an anecdote with you. The author asked the AI to pretend to be himself to chat with his mother. The AI refused his offer based on ethical issues. It is a surprising episode. The AI could handle ethical issues and choose not to obey a human instruction. Since the AI does not have a body, I think it is hard to understand the physical world. However, it also seems to be able to comprehend. The AI could tell the author how to carry a sofa to the rooftop of his house. It showed a lot of steps and explained a concrete process. The AI can do various tasks. One of the best ways to utilize it is for paperwork. Medical staff suffer from a lot of paperwork in addition to patient care. It can lead to burnout. If we use AI wisely, we can reduce the amount of paperwork. AI has downsides. There are a lot of biases to influence AI. Based on the date that it can access, it produces the output. If the data have a lot of biases, the output is not trustworthy. Therefore, we need to verify the output. It is sometimes hard because the process of producing a conclusion is invisible. We should let it show the inner workings.