Here Is How Generative AI Really Works In Healthcare – This And More News In Digital Health This Week

ChatGPT celebrated its first birthday last week. It feels like its launch was ages ago, considering how much the publicly available model advanced in the past twelve months. It has become the assistant I always longed for, and given the monumental change it brings, my research focus has also shifted toward this emerging and exciting topic. And of course, we've seen the first studies analyzing its potential.

After Lancet launched Lancet Digital Health and Nature Publishing Group launched npj Digital Medicine, now the New England Journal of Medicine also launched NEJM AI , a new medical journal meant "to identify and evaluate state-of-the-art applications of artificial intelligence to clinical medicine. Wonderful news, but I cannot wait to find out how incredibly expensive it will be to publish in that journal.

HOW EPIC IS USING AI TO CHANGE THE WAY EHRS WORK

Since Epic announced that it would tap Microsoft to accelerate generative AI-powered tools to help clinicians save time while using their electronic medical record systems, I've been curious about what it might look like in action. Now Sumit Rana, head of research and development, gave an interview about that.

"One site reported average savings of five-and-a-half hours per week. Another one looked at time spent by doctors after clinical hours, and they saw a 76% reduction in that time spent after clinic hours."

READ MORE

GPT-4 OUTPERFORMED HUMAN READERS IN DIAGNOSING COMPLEX CLINICAL CASES

"A study published in the New England Journal of Medicine says GPT-4 correctly diagnosed 52.7% of complex cases compared to only 36% of medical journal readers."

Another example of media sensationalism: as if a poll in a medical journal is equivalent to how physicians actually work in practice with a lot of responsibility. It's not. Still, it's good news for the generative AI community.

READ MORE ?

INTRODUCTION TO ARTIFICIAL INTELLIGENCE IN MEDICINE AND HEALTHCARE

Comfortably working with AI and having a scientific understanding of it is becoming a must-have skill in medicine. You don't need to become a deep-learning expert, but having a general overview will keep you up-to-date.?

We designed this course to prepare you to make the most of this revolution. You won't design neural networks, but will grasp what AI is (and isn't) good for, when to use it, when to trust it, when is it safe, and when dangerous. Come as you are, and test the first lessons for free.??

ENROLL HERE

WHAT DO WE NEED TO HAVE AI-EQUIPPED NANOBOTS IN MEDICINE

AI nanobots promise a future where diseases could become obsolete, thanks to early disease detection, treatment at a cellular level, and enhancement of physical and cognitive abilities – according to the picture painted by sci-fi writers.

We have significant progress on the micro-level, but the nanoworld is still further away, although we see some progress in areas like nanomedicine.

READ MORE

WITNESSING THE BIRTH OF MULTIMODALITY

We are witnessing the birth of multimodal large language models!

ChatGPT allows third-party tools to be integrated into the original model and it means you can now create highly realistic videos from text! I interacted with ChatGPT to prompt it to create a video for me.

READ MORE

AMA ISSUES NEW PRINCIPLES FOR AI DEVELOPMENT, DEPLOYMENT & USE

The American Medical Association (AMA) released new principles for augmented intelligence (AI) development, deployment, and use. It covers 7 key concepts.

These are: 1. Oversight, 2. Transparency, 3. Disclosure and documentation, 4. Generative AI, 5. Privacy and security, 6. Bias mitigation, and 7. Liability.

READ MORE

AI’S UNFORESEEN MEDICAL DISCOVERIES: CURIOUS UNUSUAL ASSOCIATIONS

Determining patients’ race from chest x-rays alone or diagnosing type 2 diabetes from short audio samples. AI can do it and we don’t know how. There are fascinating examples of unusual associations.

Every now and then AI has curious medical discoveries, detecting things that – to the best of our human knowledge – should not be detectable from the input data. In these cases, the medical detective work has a new aim: to understand how the AI came to the conclusion we, humans overlooked for decades.

READ MORE

MORE NEWS ABOUT THE FUTURE OF MEDICINE

DEFIBRILLATOR DRONES –? Everdrone First on Scene validated by study in Lancet Digital Health

12.7% REDUCTION? –? AI chatbot reduces depression in prenatal and postnatal women

DELICATE BALANCE – Generative AI could revolutionize health care — but not if control is ceded to big tech

HELLO ANTHROBOTS – Tiny living robots made from human cells surprise scientists