As part of IA index report issued by the Institute of Artificial Intelligence Centered on Man (HAI), experts detailed artificial intelligence trends in the past year and how it transforms society. This year, the report presents an extended chapter on science and medicine, developed by a team of Make healthA collaboration between Stanford Medicine and Hai.
Russ AltmanMD, PHD, professor of bio-engineering, genetics and science of biomedical data, helped direct the development of the chapter of sciences and medicine, which highlighted AI milestones, such as progress in protein and molecules, support for clinical care and automated detection of diseases.
Altman discussed the best dishes to remember of the year, how the trends of the AI will shape the future of science and biomedical medicine, and what it considers the most promising areas of growth.
The section of this report on science and medicine considerably expands past years. How does it reflect what you see with regard to the EA explosion?
The section of last year did an excellent job by highlighting certain key examples of AI in science and medicine, almost as a biopsy of the situation as a whole. We have something called a systems exam when we see a patient: you make a full assessment to make sure you don’t miss anything – assess the skin, eyes, heart, lungs, etc. We believe that this year’s chapter has gone from a biopsy, just looking at a small place here and there, to an examination of the systems.
Anyone at the Faculty of Medicine knows that AI is everywhere. You could have lunch with someone who has never struck you as an AI researcher, and you discover that he is doing an effort to build a large language model in his clinical practice. I host a podcast, The future of everythingAnd I see the same trend there. I question teachers from all over the university; The most common thing they tell me just before pressing the file is: “Make sure you ask me how AI is revolutionizing my work.” It really revolutionizes university life and scholarships, and it’s quite exciting.
The chapter of science and medicine addresses a variety of trends, the impact of AI on clinical care on ethical considerations surrounding the use and development of AI. What were your three biggest dishes to take away?
The first is the creation of foundation models. A foundation model is essentially a statistical model which describes a very large set of data. About 10 to 15 years ago, we all talked about “Big Data”: people collected huge amounts of data, but we did not always know what to do with it, and in many cases it was too much. Scientists would end up picking cherries, using clean data and points to well -stored learning. The foundation models allow us to look at all the cherries, not only those that are ripe and perfect and at the eye level. The foundation models take all the data from your large data set and put it towards a rich statistical model so that it can make predictions and projections. This is why so many sciences see such rapid progress – because scientists now have a model that essentially allows them to speak to their data, ask a question and get an answer.
Second, AI research has contributed directly to two Nobel prizes. It is a real interest in the soil. Yes, there is a media threshing, and there are questions to find out if AI is good or bad for society. For science, it’s good, and I cannot think of a better short -lived validation than “two Nobel prizes with AI technology were awarded the same year.” This is the title I would say to my mother or children: “Yes, it’s real.”
Third, the ability for us to use large language models to improve all parties of clinical care is enormous. Many of my clinical colleagues who are in the trenches every day are interested in integrating AI into their daily work flow, such as using an LLM to help them write notes, or to listen and watch a surgery, then receive a quality summary of what happened in the operating room. Great languages can reduce what is called “pajamas time” – the hours that doctors spend after closing the clinic, catching all their papers. This can have a great negative effect on the quality of life.
Where do you see the most potential for AI during the coming year?
The ability of large language models to deliver messages at different levels of education or with nuances of different cultural backgrounds is an enormous unexploited opportunity. Linguistic models can distill information to help patients understand their disease and treatment plan. AI can suggest means to communicate effectively or provide different perspectives that the doctor may not have thought.
For example, someone can be opposed to taking pills. You can imagine that the chatbot leads the doctor to say something like: “I hear you. I understand that you are not a person in pill. There are five tablets that could have been prescribed, but we only give you two, because these are the most important. ” I am optimistic that better communication and clarity will lead to an improvement in the understanding of patients in their disease and therefore to improve the therapeutic doctor-patient.
This story was reported for the first time by Stanford Medicine.