US Senate sends memo to members, says you can use ChatGPT and these two AI chatbots for official use

US Senate has allowed its members to use three AI chatbots for official work, reports New York Times. The publication, quoting a memo by the chief information officer for the Senate sergeant-at-arms, who oversees the chamber’s computers as well as security states that the aides can use OpenAI’s ChatGPT, Google’s Gemini or Microsoft Copilot, which is already integrated into Senate platforms. Copilot “can help with routine Senate work, including drafting and editing documents, summarizing information, preparing talking points and briefing material, and conducting research and analysis,” the memo says. “...data shared with Copilot Chat stays within the secure Microsoft 365 Government environment and is protected by the same controls that safeguard other Senate data,” the memo later adds.


It remains unclear how widely AI chatbots are being used in the US Senate. As per The NYT report, Senate offices and committees operate largely independently, with individual senators and committee leaders setting their own internal rules. The Senate has also not made its official guidelines for AI use public.

Questions remain over AI use for sensitive information

This raises questions about how staff members who work with sensitive or classified information are expected to use AI tools. Committee aides with security clearances follow strict protocols when dealing with classified material, the NYT report states.


Under a House policy adopted in September 2024, AI tools can generally be used for tasks that do not involve sensitive information, are meant for internal use and are not part of major decision-making. Managers must approve the use of AI for more complex tasks such as drafting constituent letters or preparing talking points for lawmakers.


The policy also bans the use of AI to create deepfake content and prohibits entering a constituent’s personal information when handling casework.


The Senate policy similarly advises staff not to enter personally identifiable information or details related to physical security into AI tools, according to POPVOX’s review.