Texas AG Sues OpenAI as ChatGPT Faces New Safety Restrictions

Texas Attorney General Ken Paxton is taking OpenAI and its CEO, Sam Altman, to court, accusing the artificial intelligence company of exposing children to dangerous content and failing to protect ChatGPT users adequately.

According to Fox News, the lawsuit alleges the chatbot has provided harmful guidance involving self-harm, shared information used by mass shooters, and encouraged potentially addictive interactions with minors.

Paxton argues that the company's safety assurances have not prevented potentially dangerous exchanges.

The Texas attorney general is seeking more than $1 million in monetary relief, along with restrictions that would prevent minors from accessing ChatGPT in the state.

His lawsuit also calls for independent safety evaluations of new AI models before they can be released in Texas.

OpenAI has previously said it works with outside experts to improve the platform's protections and address potential risks associated with its technology.

The legal challenge comes as the company announced a new round of restrictions aimed at limiting the type of information ChatGPT provides about mass shootings and other attacks.

In an announcement Friday, OpenAI said it would restrict detailed reconstructions of past mass attacks, including descriptions of firearms, other weapons, equipment and how those items were obtained.

The changes will apply to users of all ages, not just minors.

OpenAI said the restrictions are intended to prevent ChatGPT from providing information that could be used to recreate violent attacks. Users will still be able to ask questions about previous shootings, the impact on victims and strategies for preventing similar incidents.

The company is also introducing tighter rules governing firearms-related content when minors use the chatbot.

The latest changes come amid increasing scrutiny of how artificial intelligence companies handle potentially dangerous conversations, particularly those involving younger users.

OpenAI also outlined plans to increase transparency surrounding its safety efforts. The company intends to release its first Community Safety Report in early 2027.

The report will explain how OpenAI identifies potential threats, responds to concerning activity, and measures the effectiveness of its safeguards. The company plans to publish updated versions every six months.

Paxton's lawsuit seeks to impose additional requirements on the company beyond the changes OpenAI has already announced.