Artificial intelligence is inserting itself into our lives in various ways. New technology always has its pros and cons—but while the cons of using AI can be harrowing for individuals prone to psychotic episodes, it can also be used to predict when and how a person with psychosis will respond to interventions designed to help.

Large Language Models (LLMs) like ChatGPT are a type of machine learning that uses human language trained on many other linguistic datasets to give results and answers in the same format. LLMs can often mimic or appear to be human-like because it takes the voice of a human, potentially leading the user to believe they are truly talking to a person.

The type of machine learning used in scientific studies is different than LLMs. Machine learning used in mathematical or scientific fields takes data (demographic data, measurements, test scores, and more) and gives predictions based on a desired goal the user prompts it with, displayed as numerical information rather than linguistic. Used in this way, machine learning can be a helpful tool to make predictions about what can be done to help people. Those predictive models are typically focused on preventive and positive goals, like preventing relapses—or perhaps someday, predicting who will get psychosis in the first place.

Right now, we are still........

© Psychology Today