Advances in artificial intelligence technology are rapidly changing every industry, including healthcare. Generative AI is the latest leap in innovation, exemplified by chatbots like ChatGPT that are now widely available. But what does generative AI mean for doctors?

Technology changes fast, and keeping up takes a lot of work. If you’re a physician or medical professional not already well-versed in generative AI, this article will provide a brief introduction in simple terms. We’ll explain generative AI, examine how it’s already being used, and point you toward resources for further learning.

What is generative AI?

Generative AI is a technology that relies on deep-learning algorithms to create new content like text, audio, code, and images. Generative AI is a form of machine learning, a larger category of techniques that allow computers to use algorithms to learn from data.

Machine learning has been used widely in medicine for several years, for example, to automate diagnostic imaging or predict outcomes based on patient information. In these cases, computers are fed large amounts of data and can observe and clarify patterns in valuable ways.

Like all types of machine learning, generative AI uses algorithms to learn from input data. But it goes further to create new content based on past learning – hence the term “generative.” This ability to create new outputs sets generative AI apart from traditional AI.

Take dermatology as an example. With traditional AI, you could program a system to recognize specific visual features associated with a particular skin disease. After reviewing thousands of images, the system learns to provide a reliable diagnosis. This process has led to remarkable advancements in automated skin disease classifications in recent years.

In contrast to traditional AI, a generative AI system could be trained on a wide variety of skin images, learning subtle patterns and variations that developers did not program explicitly. When faced with a new skin image, the program wouldn’t just recognize known conditions but might also identify emerging or rare skin diseases that were not part of its initial programming.

Generative AI’s ability to create new data similar to input or training data opens up a whole new range of applications in healthcare and other industries. 

You already use generative AI in healthcare

Every month, more of the technology we use incorporates generative AI. AI’s presence may not always be noticeable, but it increasingly powers everything from medical imaging systems to workflow optimization.

Here are a few examples of how generative AI is already deployed in everyday healthcare settings physicians might encounter:

  • Chatbots and virtual health assistants – Many hospital and clinical websites and mobile apps now have chatbots or “virtual health assistants” that can answer common questions, provide basic health information, and even schedule appointments. These chatbots rely on generative AI to simulate a realistic question-and-answer exchange that frees up time for administrative and clinical staff. 
  • Workflow optimization – Large hospitals use software to optimize the flow of people and resources. By analyzing historical data, generative AI can produce insights that guide patient admission rates, staffing needs, and resource utilization. 
  • Clinical decision support – Generative AI can enhance clinical decision support systems by offering insights based on patient data, medical literature, and best practices. Some surveys show clinicians already use generative AI tools to assist with about 1 in 10 clinical decisions.
  • Claims processing – Health insurance prior authorization and claims processing are time-intensive and costly. Payers are starting to use generative AI to verify benefits, calculate out-of-pocket costs, consolidate codes, draft response letters, and much more.

Emerging use cases

There are several healthcare challenges that generative AI could solve in a big way, such as improving patient monitoring, early detection of cancers, and accelerating clinical trials. All of these are exciting, but there’s one area in which generative AI will transform healthcare providers’ work on a day-to-day level: note-taking.

Time spent on charting and EHR data entry is a significant source of frustration for many physicians. The good news is that software is emerging to automate clinical note-taking using generative AI and natural language processing.

Here’s how it works: a clinical records a patient visit using an AI-enabled mobile app. The software transcribes the interaction in real-time, much like a scribe would taking a dictation. Once the visit ends, the AI generates a note in a pre-structured format corresponding to electronic health record fields. The clinician reviews the note on a computer, makes any edits, and submits it to the EHR.

While not yet widely used by physicians and nurses, more companies are releasing software that automates note-taking with remarkable accuracy. These generative AI tools eliminate about 80% of the administrative time associated with every patient interaction, freeing providers to see more patients and focus on the fulfilling part of their work.

Generative AI resources for doctors

There are several new initiatives for healthcare providers interested in learning more about AI.

The Health AI Partnership (HAIP) is a multi-stakeholder network that empowers healthcare organizations to use AI safely, effectively, and equitably. HAIP is creating excellent resources, including a guide with 8 key decision points to consider when implementing an AI solution

The American Medical Association Augmented Intelligence page includes the latest AI news, resources about CPT and AI, and articles about AI and practice management, policy, and more.

Similarly, JAMA and other leading medical publications provide articles and resources on artificial intelligence and large language models. These are great places to start if you want to deepen your understanding of AI or incorporate generative AI tools in your clinical practice. 

Comments are closed.