The U.S. House of Representatives is setting new boundaries for the use of AI chatbots, specifically ChatGPT, within congressional offices. This move underscores the increasing concern in Washington about the implications of the rapid expansion of generative AI.
The House’s Chief Administrative Officer, Catherine L. Szpindor, issued a memo stating that offices are “only authorized” to use the paid version of ChatGPT Plus. This version, unlike the free service, incorporates critical privacy features necessary to safeguard House data. The memo also explicitly stated that no other large language models are authorized for use.
The guidelines for using the tool are clear: it can be used for “research and evaluation only,” and offices are not authorized to incorporate it into their regular workflow. Staffers are instructed to input only “non-sensitive” data and avoid pasting any text that hasn’t already been made public. Privacy settings, which are disabled by default, must be enabled to ensure that interaction history is not preserved and not incorporated back into the large language model.
This cautionary approach comes as lawmakers from both sides of the aisle are rushing to draft legislation regulating AI. A bipartisan group of senators, including Senate Majority Leader Chuck Schumer, is working on comprehensive legislation. Meanwhile, Senators Richard Blumenthal and Josh Hawley, who lead a Senate Judiciary subcommittee on tech, have held hearings on the issue.
Congress, like other institutions, is grappling with how generative AI fits into its workflow while also wrestling with broader questions about the technology. If you want to help your team become more productive (and safer) using ChatGPT Plus, have them sign up for our free online course, Generative AI for Execs.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it.