AI is being applied in research, administrative workflows, clinical care, and more. From automated data processing to intelligent tutoring systems, AI supports innovation across multiple disciplines.

AI tools are available to students, faculty, and staff. Some require approval based on usage needs and data security policies.

Generative AI refers to a type of artificial intelligence that can generate new data that resembles a given set of training data. These systems can produce anything from images and music to texts and videos.

Common uses include generation of custom images, videos, text, and code. In addition, generative AI can augment data, create music, and even design a video game.

Most generative models, like Generative Adversarial Networks (GANs)* or Variational Autoencoders (VAEs), learn to represent the underlying patterns in the data they're trained on. They then use this learned representation to generate new, similar data.

*A Generative Adversarial Network (GAN) is a type of neural network architecture where two networks, a generator and a discriminator, are trained together. The generator tries to produce fake data, while the discriminator attempts to distinguish between real and fake data. Over time, the generator improves its ability to generate convincing data.

While generative AI aims to produce novel outputs, there's no guarantee that they are always unique. However, the vast potential output space means that there's a high likelihood of generating something different each time.

While generative AI aims to produce novel outputs, there's no guarantee that they are always unique. However, the vast potential output space means that there's a high likelihood of generating something different each time.

Traditional AI (often rule-based or supervised learning) typically requires explicit programming or labeled data to produce a specific output. In contrast, generative AI can produce new, previously unseen content based on patterns it has learned.

Yes. There's the potential for misuse in generating misleading content, fake news, or deepfakes*.

In addition, as shared by the University in May 2023, the generative AI tool ChatGPT is not HIPAA compliant, and the University does not permit its use with any patient data. OpenAI, the company behind ChatGPT, also warns against inputting confidential data into the platform, which could constitute a public disclosure and lead to a loss in our ability to protect University information, including intellectual property.

Important: Copilot and Gemini are not HIPAA compliant yet. Copilot and Gemini unlocks generative AI for capabilities around asking questions and generating responses from the web. These services do, however, have some limitations, such as conducting custom searches within University of Miami firewalls. If you need any assistance around similar examples, or other potential use cases surrounding generative AI, please reach out to the Department of Data Science and Research Informatics at dsri@med.miami.edu or Academic Technologies at academictechnologies@miami.edu.

*A deepfake is a synthetic media (typically video) in which a person's likeness is replaced with another's, making it appear that the person is saying or doing something they didn't. This is achieved using advanced generative AI techniques.