Musk’s DOGE Used Meta’s Llama 2, Not Grok, for Gov’t Cuts: Report
▼ Summary
– An outdated Meta AI model, Llama 2, was used by Elon Musk-affiliated DOGE to review federal workers’ responses to the “Fork in the Road” memo, which mirrored Musk’s Twitter ultimatum.
– The “Fork in the Road” memo gave federal employees the choice to accept return-to-office policies or resign, and Llama 2 helped classify responses to track resignations.
– Llama 2 was previously involved in a scandal when Chinese researchers used it for a military AI model, prompting Meta to reverse its ban on military applications.
– Meta announced partnerships with US government agencies and private firms like Lockheed Martin and Microsoft to provide Llama for national security and defense uses.
– Wired noted that Meta’s open-source models allow government use without explicit consent, enabling alignment with Musk’s objectives.
Reports indicate that an older Meta AI system played a key role in the Department of Government Efficiency’s controversial workforce reduction strategy. Internal documents reviewed by Wired reveal that teams linked to Elon Musk’s DOGE initiative leveraged Meta’s Llama 2 model to analyze responses from federal employees following the widely circulated “Fork in the Road” email.
The memo, which mirrored Musk’s earlier ultimatum to Twitter staff, forced government workers to either commit to stricter office policies or leave their jobs. While speculation swirled about AI-driven workforce evaluations, newly uncovered records confirm Llama 2 was deployed to categorize resignation notices and gauge employee pushback.
This isn’t the first time Llama 2 has been embroiled in controversy. Late last year, Chinese military researchers reportedly built an AI system using the model, prompting Meta to publicly distance itself from the project. However, the company later revised its stance, opening access to U.S. defense and national security agencies alongside major tech partners like Microsoft, Oracle, and Palantir.
A Meta spokesperson emphasized the shift, stating the company now actively supports government applications through collaborations with firms specializing in defense and data analytics. Because Llama 2 operates as open-source software, agencies can adapt it without direct oversight—a flexibility that reportedly allowed DOGE to integrate the tool without Meta’s direct involvement.
The revelation raises questions about how widely AI is being used to streamline—or disrupt—federal operations, particularly when paired with aggressive workforce policies. While Meta has positioned itself as a willing partner for official projects, the unintended use of its technology highlights the challenges of regulating open-source AI in high-stakes environments.
(Source: Ars Technica)
