U.S. House Restricts Use of Large AI Language Models

Except for OpenAI’s ChatGPT Plus service, the United States House of Representatives recently adopted new rules prohibiting the use of large artificial intelligence (AI) language models by members.

According to a notice sent by Catherine Szpindor, the chief administrative officer of the U.S. House, the use of chatbots has been restricted for security reasons:

“House offices using the ChatGPT artificial intelligence chatbot are only authorized to use the ChatGPT Plus version of the product. The Plus version of the product incorporates important privacy features that are necessary to protect House data.”

No other versions of ChatGPT or other large language models of Al software are currently authorized for use in the House, according to the memo.

The document also contains provisions limiting House members’ use of the software to “research and evaluation only” and prohibiting the integration of ChatGPT into their “regular workflow.”

Other restrictions include restrictions on sharing sensitive data as notifications and a requirement that all ChatGPT Plus privacy settings be enabled.

OpenAI has not listed any privacy-related benefits exclusive to the Plus service, so it is unclear precisely what privacy features the document refers to.

According to OpenAI, ChatGPT Plus provides public access to the model during peak hours, quicker query response times, and priority access to new features. No mention is made of additional privacy features.

In April, the company added the ability for ChatGPT and ChatGPT Plus users to erase their chat history and accounts; however, information deleted in this manner remains on the ChatGPT servers for an additional 30 days.

According to a blog post published in April, OpenAI intends to launch a ChatGPT business subscription service with enhanced data control capabilities. There is no indication of how this service may differ from ChatGPT Plus.

Once adopted, the new House rules will only apply directly to House members; however, U.S. Representatives Ted Lieu, Ken Buck, and Anna Eshoo recently introduced a bipartisan bill that would establish a federal artificial intelligence commission with the mandate to provide regulatory oversight to the U.S. AI industry as a whole.