344
Productivity & Workflow355
Automation & Workflow224
Software Development251
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative170
Photography & Imaging156
Customer Support131
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Documents show HUD withheld records about DOGE’s AI use in housing policy, including by citing a privilege that legal experts say does not exist.
In short: HUD has withheld more than 100 documents about DOGE’s use of AI at the housing agency, and it cited an “AI privilege” that experts say is not real.
A team linked to the so called Department of Government Efficiency, or DOGE, worked inside the US Department of Housing and Urban Development, or HUD. According to documents released through a public records process, the team used AI to help review HUD rules and inform policy decisions.
Democracy Forward, a nonprofit legal group, filed a Freedom of Information Act request, often called FOIA. FOIA is a law that lets the public ask for government records. HUD withheld more than 100 requested documents about how AI tools were developed and used.
Many of the withheld files have names that suggest AI was part of the work, including documents about “prompts.” A prompt is the text someone types into an AI chatbot, like giving instructions to a very fast assistant. HUD’s FOIA office cited the usual “deliberative process” exemption in many cases, which is meant to protect internal back and forth before a decision is final.
However, in some cases HUD cited reasons such as “draft of AI prompt,” “deliberative AI input,” and an “AI privilege.” Privacy and transparency experts quoted by WIRED say FOIA does not include any special exemption for AI, and that computers are not entitled to the same protections meant for human discussion.
If AI helps shape housing rules, the public may want to know what questions officials asked it and how much they relied on the answers, especially since AI systems can make things up or get facts wrong (like a confident assistant who sometimes guesses). There is currently no US law that requires agencies to say when AI was used to create or change policies.
Source: Wired