that the project is powered by Elon Musk’s artificial intelligence company, xAI.
The chatbot is publicly available for use and is hosted on a DOGE-named subdomain on Christopher Stanley's website. Stanley is the head of security engineering for SpaceX and the White House.
However, it was not mentioned whether or not the chatbot is experimental or whether it was used by DOGE as part of theri not foreseen cost-cutting efforts across the U.S. government that led to legal and privacy concerns.
Christopher Stanley along with a White House Spokesperson have not responded to a request for comment as of this time.
It is also worth mentioning that the chatbot names itself “ Department of Government Efficiency AI Assistant” and lets the users know that it is powered by xAI’s Grok-2. The chatbot also said that it is here to “assist government personnel in identifying waste and improving efficiency.”, reported TechCrunch.
The AI chatbot seems to work as a large language model in that it is trained on different key DOGE goals some of which have the goal of creating a “less dumb” and deleting “unnecessary parts or processes”.
An example of how the artificial intelligence chatbot works was reported by TechCrunch which asked the AI what DOGE should do about USAID which stands for a federal agency that was shuttered by DOGE’s reforms, it replied with its five principles also recommending to eliminate any “bureaucratic layers” when it comes to USAID and decision-makers.
Even more so, when asked which political figures he should imitate he replied that using the five principles, mentioning the former UK prime minister Margaret Thatcher and former prime minister from Singapore Lee Kuan Yew, also adding excellent models for DOGE” by searching for “efficiency, simplification, and the use of technology.”.
The chatbot has some issues that are usually common for every large language model – hallucination. Sometimes also give wrong answers or try not to answer some of the questions.
However, it is not clear whether Musk's use of xAI would constitute a conflict of interest. Since LLM usually charges users through API usage, government workers using a chatbot powered by xAI could increase xAI’s revenue.
.webp)
