Skip to evidence
Tiny Red HandsSearch
Back

DOGE Used a Meta AI Model to Review Emails From Federal Workers

WIRED

Read original source

Reporting cited here

  1. reported

    WIRED reported that DOGE used Meta's Llama 2 large-language model to review responses to the recurring federal-worker accomplishment email, while the role of model output in personnel decisions remained unclear.

    Passages and source records documenting: DOGE used Meta's Llama 2 model to analyze federal workers' email responses

    DOGE used Meta's Llama 2 model to analyze federal workers' email responses · OPM acknowledged automated processing while withholding detailed decision rules