Ravi 
In a bold move that underscores the growing integration of artificial intelligence in government operations, Elon Musk’s Department of Government Efficiency (DOGE) has deployed a custom chatbot called GSAi to 1,500 federal workers at the General Services Administration (GSA). This strategic implementation signals a significant shift in how government technology teams approach daily tasks and workforce efficiency.
The Rise of GSAi: A New AI Tool for Federal Workers
Developed over several months and initially tested with 150 users in February, GSAi represents a cutting-edge approach to workplace automation. The chatbot offers federal employees a versatile AI assistant capable of:
- Drafting emails
- Creating talking points
- Summarizing text
- Writing code
Flexible AI Model Selection
One unique feature of GSAi is its flexibility in AI model selection. Users can choose between different models, including:
- Claude Haiku 3.5 (default)
- Claude Sonnet 3.5 v2
- Meta LLaMa 3.2
Strict Guidelines and Limitations
Despite its capabilities, GSAi comes with stringent usage guidelines. Employees are explicitly instructed not to input:
- Non-public federal information
- Pre-decisional documents
- Internal communications
- Personally identifiable information
Workforce Implications
The chatbot’s deployment coincides with significant workforce reductions at the Technology Transformation Services (TTS). Thomas Shedd, a former Tesla engineer now running TTS, announced plans to shrink the tech branch by 50%, emphasizing a “results-oriented and high-performance team” approach.
Employee Perspectives
Initial user feedback suggests a pragmatic view of the technology. One employee described GSAi as “about as good as an intern” – indicating potential but also acknowledging current limitations.
Future Outlook
While GSAi’s current deployment is limited to the GSA, there are indications of broader interest. The Treasury and Department of Health and Human Services have reportedly considered similar chatbot implementations.
Source: Original reporting by Makena Kelly and Zoë Schiffer, WIRED
