LLMs are also coming to healthcare, and indeed it was only a matter of time before that happened. We are talking about Large Language Models based on neural networks, which make possible artificial intelligence technologies such as ChatGTP, Bard, Bing Chat and other solutions under development. In this regard, Microsoft and Epic Systems announced that they are introducing the next generation LLM powered by OpenAI, GPT-4, as an innovator in healthcare. It will serve healthcare professionals in various types of uses, such as the analysis of medical records and the search for information on a large scale.
Epic Systems is one of the largest healthcare software companies overseas. Its electronic health record software is reportedly used in more than 29 percent of acute care hospitals in the United States, and more than 305 million patients have an electronic health record in Epic worldwide. Epic’s use of predictive algorithms in healthcare has come under criticism in the past.
In the announcement, Microsoft mentions two specific ways Epic will use its Azure OpenAI service, namely one that provides API access to OpenAI’s large language models (LLMs), such as GPT-3 and GPT-4.
The first use of GPT-4 will allow doctors and operators to create automatic replies to patient messages, relieving them of tasks considered minor, in favor of a more prudent use of time. The communiqué is quite clear in this.
“Integrating Generative AI into some of our day-to-day workflows will increase the productivity of many of our providers, allowing them to focus on the clinical tasks that truly need their attention.”
The second use is also similar to what many can experience with chatbots, i.e. the ability to ask questions in text form about any content in the databases and request any type of data analysis. For example, this tool could allow you to research large numbers of patients to identify trends that could be useful for both medical research and financial reasons. Microsoft gives some very intuitive examples and apparently it will be possible to interrogate the system as is done with any chatbot, but to have detailed medical information about one or more patients, or trends related to a disease, and much more.
The news is causing a sensation and as you can imagine there is a dispute underway. On the one hand there are the supporters of artificial intelligence technologies, who see a bright future in LLMs, able in this case to give real added value in the ability to diagnose, treat, and research in the medical field.
On the other hand, there are those who say the GPT-4 technology is still immature and face the risk of erroneous data, unreliable results, or even discrimination based on gender, race, age, or other factors. The question is quite complex therefore we are sure that we will talk about it again in the future.
In the meantime, Europe is discussing how to regulate ChatGPT and AI technologies after the ‘LA’ given by Italy on the issue. In this regard, we remind you that the Privacy Guarantor has confirmed to the creators of ChatGPT that the popular chatbot will be able to function again without risk in Italy as long as it complies with a series of rules by April 30th. Which? We talked about it in detail in this article.