Congress Embraces AI Tools Amid Unclear Rules and Oversight
Lawmakers and their staff are increasingly using AI chatbots for drafting legislation and communications, but many are unsure of the rules governing their use.
Congressional staff across the Senate and House have been granted permission to employ AI platforms—including ChatGPT, Microsoft’s Copilot, Google’s Gemini and Anthropic’s Claude—for a variety of duties such as drafting speeches, preparing hearing questions and even writing bill amendments. The House purchased 6,000 Copilot licenses last year, encouraging widespread adoption, and several lawmakers, like Senator Tommy Tuberville, have described using ChatGPT while traveling to inform their positions on issues like Medicare.
Despite this integration, many staff members are uncertain about the specific restrictions, which forbid using AI for deepfakes, personnel decisions, handling sensitive constituent information, or finalizing legislation. Enforcement is informal, relying on individual vigilance, and confusion has already led to mishaps, such as a chatbot-generated line appearing in an amendment proposed by Representative Anna Paulina Luna. As AI use expands without clear oversight, the risk of further procedural errors and privacy breaches grows.
Why it matters
Unregulated AI use in Congress could lead to policy mistakes and privacy violations affecting the public.
In this story
