# AI-Generated Images of Substance Use and Recovery: Mixed Methods Case Study

## TL;DR

This study explores how AI-generated images of substance use and recovery can be stigmatizing and biased, and how guidelines can help reduce these issues.

## Contribution

The study introduces a mixed-methods approach to evaluate and improve AI-generated images related to substance use disorder using guidelines and custom prompts.

## Key findings

- Default AI-generated images predominantly depicted White men and contained stigmatizing elements like injection drug use and dark colors.
- Guideline-informed custom prompts produced less stigmatizing images but mostly depicted Black women.
- Clear guidelines can improve AI-generated images, but further iteration is needed to fully align with best practices.

## Abstract

Images created with generative artificial intelligence (AI) tools are increasingly used for health communication due to their ease of use, speed, accessibility, and low cost. However, AI-generated images may bring practical and ethical risks to health practitioners and the public, including through the perpetuation of stigma against vulnerable and historically marginalized groups.

To understand the potential value of AI-generated images for health care and public health communication, we sought to analyze images of substance use disorder and recovery generated with ChatGPT. Specifically, we sought to investigate: (1) the default visual outputs produced in response to a range of prompts about substance use disorder and recovery, and (2) the extent to which prompt modification and guideline-informed prompting could mitigate potentially stigmatizing imagery.

We performed a mixed-methods case study examining depictions of substance use and recovery in images generated by ChatGPT 4.o. We generated images (n=84) using (1) prompts with colloquial and stigmatizing language, (2) prompts that follow best practices for person-first language, (3) image prompts written by ChatGPT, and (4) a custom GPT informed by guidelines for images of SUD. We then used a mixed-methods approach to analyze images for demographics and stigmatizing elements.

Images produced in the default ChatGPT model featured primarily White men (81%, n=34). Further, images tended to be stigmatizing, featuring injection drug use, dark colors, and symbolic elements such as chains. These trends persisted even when person-first language prompts were used. Images produced by the guideline-informed custom GPT were markedly less stigmatizing; however, they featured almost only Black women (74%, n=31).

Our findings confirm prior research about stigma and biases in AI-generated images and extend this literature to substance use. However, our findings also suggest that (1) images can be improved when clear guidelines are provided and (2) even with guidelines, iteration is needed to create an image that fully concords with best practices.