The US House of Representatives has taken a firm stance on data security by prohibiting congressional staff from using Microsoft’s AI-based chatbot, Copilot, on government-issued devices. This decision, made by the House’s Chief Administrative Officer Catherine Szpindor, comes in response to concerns over the potential leakage of sensitive information to unauthorised cloud services.
In a memo obtained by Axios, Szpindor stated that Copilot would be removed and blocked on all House Windows devices, following a similar restriction imposed last June on the use of ChatGPT, another AI-based chatbot. The Office of Cybersecurity has deemed Copilot a risk to users due to the threat of leaking House data to non-approved cloud services.
Microsoft has acknowledged the higher security requirements of government users and is working on developing a roadmap of AI tools, including Copilot, that meet federal government security and compliance requirements. The company plans to deliver these tools later this year.
The Chief Administrative Officer’s office clarified that the current guidance applies to the commercial version of Copilot, and they will evaluate the government version once it becomes available to determine its suitability for use on House devices.
This move underscores the House’s commitment to safeguarding sensitive information and follows a similar restriction imposed last June on the use of ChatGPT. Staffers were allowed limited use of the paid subscription version of ChatGPT while the free version was banned altogether.