Accurate, Focused Research on Law, Technology and Knowledge Discovery Since 2002

Using AI to Comply With Book Bans Makes Those Laws More Dangerous

Brennan Center for Justice: “In August, a public school district in Iowa reportedly used ChatGPT to help it comply with the state’s controversial book ban law. That law—like counterparts passed in Florida, Texas, Missouri, Utah, and South Carolina—seeks to limit discussion of gender identity and sexuality in schools by barring school libraries from carrying books that touch on topics related to sexual content. Driven by a growing number of organized groups and political pressure from state lawmakers, such laws threaten students’ rights to access information and run the risk of chilling speech on topics such as sexuality, teen pregnancy, and sexual health. And using generative AI tools to implement these laws only compounds the problem. Widely accessible generative AI chatbots like OpenAI’s ChatGPT and Google’s Bard provide individuals free access to tools built on large language models. Earlier language models and other analogous automated tools have been used by businesses to detect spam, recommend products, and moderate online content. Years of experience in content moderation shows that these tools are simply not suited to making decisions based on the types of vague and subjective standards that characterize many book bans. Even for the narrower category of sexual content, AI tools are often overinclusive…”

Sorry, comments are closed for this post.