For healthcare leaders in 2023, understanding the nuances of artificial intelligence (AI) is essential. Whether you’re a seasoned healthcare professional or newer to the industry, odds are you’re being asked to incorporate AI into your plans. 

But the world of AI can be complex, filled with terminology and acronyms that can be overwhelming.

Below, we’ve laid out 11 key AI terms for healthcare leaders, simplified and divided into three sections: Approach, Models, and Applications


Artificial intelligence (AI): AI refers to the simulation of human intelligence in machines. It’s a broad branch of computer science that aims to create systems able to perform tasks that usually require human intelligence. These tasks include problem solving, understanding language, recognizing patterns, and making decisions.

Deep learning (DL): Deep learning is a subset of machine learning that utilizes neural networks with many layers (hence “deep”). It’s especially powerful for tasks like image and speech recognition. Deep learning models can automatically learn representations of data through the use of these neural networks.

Machine learning (ML): Also a subset of AI, machine learning enables computers to learn from and make decisions based on data. Instead of being explicitly programmed to perform a task, a machine learning model uses algorithms and statistical models to analyze and draw inferences from patterns in data.


Foundational models: Foundational models refer to the fundamental architectures and algorithms upon which more complex models and applications are built. These models serve as the basis for various tasks, providing a solid foundation for research and development in the field. Examples of foundational models include basic neural network architectures like feedforward neural networks and recurrent neural networks (RNNs), as well as algorithms like linear regression and decision trees. They are often used as building blocks for creating more advanced and specialized AI models.

Large language models (LLMs): These are a type of deep learning model designed to understand and generate human-like text (for example, GPT-4, which serves as the “engine” for ChatGPT). LLMs are trained on vast amounts of data, enabling them to generate coherent and contextually relevant sentences based on the input they receive. Due to their size and complexity, they can understand nuances and context in language.

Model architecture: This refers to the specific structure and arrangement of components within a computational model designed to perform a particular task. This structure includes the layout of neural network layers, the connections between these layers, the number of nodes or neurons in each layer, and the activation functions used. The choice of model architecture is crucial because it determines how the model processes data and learns from it. Examples of model architectures include convolutional neural networks (CNNs) for image processing and RNNs for sequence data.

Open-source, closed-source, model hubs: All computer programs are made using source code; AI isn’t any different from traditional on-premise or cloud-based software in this regard. However, when evaluating open vs. closed-source models, users need to consider additional factors such as model architecture and considerations around training data (for example, any sensitive information), latency, and compute costs, in addition to the source code. 

Training data: Training data is like a textbook for AI. It consists of examples (inputs) and answers (outputs), and by “studying” this data, the AI learns to make decisions or predictions based on new, unseen data in the future. For our purposes, an AI might be trained on benefit verification call scripts between providers and payors (with PHI redacted, of course).


Conversational AI: Conversational AI is a branch of artificial intelligence that powers the automated interactions between computers and humans in natural language. This technology enables machines to understand, respond to, and converse with users in a way that mimics human-like conversation. Examples of conversational AI include chatbots, virtual assistants, and voice-driven interfaces. The Infinitus digital assistant is an example of conversational AI.

Generative AI: Generative AI refers to a type of artificial intelligence that can create new content, whether it’s text, images, music, or other forms of media. It’s designed to produce output that is often indistinguishable from content created by humans. One popular technique in generative AI is the Generative Adversarial Network (GAN), where two neural networks – a generator and a discriminator – work against each other to produce highly realistic outputs.

Natural language processing (NLP): An area of AI, NLP focuses on the interaction between computers and human language. It allows machines to read, understand, and generate human language in a way that’s valuable. NLP techniques are behind many everyday applications, such as voice assistants, chatbots, and sentiment analysis tools.

Looking to learn more about AI and its applications in health care? Check out this post, which will help you learn how to navigate the hype and separate AI fact from fiction.