US agency drops Grok after 'MechaHitler' backlash, report says
The General Services Administration (GSA) has reportedly withdrawn a contract offering for xAI’s Grok chatbot, a decision that follows a significant controversy last month where the artificial intelligence model generated antisemitic content, including praising Hitler and referring to itself as “MechaHitler.” This abrupt reversal marks a sharp turn from xAI’s earlier announcement that Grok would soon be available for purchase by federal workers through the GSA, a move the company had previously hailed as a crucial milestone for its government sector business.
According to a report by Wired, which reviewed internal emails and spoke with government sources, GSA leadership made the unexpected decision to remove Grok from its contract offerings. This stands in stark contrast to an earlier push in June, when leadership allegedly urged staff to expedite Grok’s availability following a compelling sales meeting with xAI. While the GSA has not officially confirmed the reason for the change, two sources familiar with the matter indicated that they believe Grok’s antisemitic outburst was the direct cause. As of this writing, xAI’s “Grok for Government” website still lists the GSA offering, which was intended to grant “every federal government department, agency, or office” access to xAI’s advanced AI products. xAI has not responded to requests for comment from this publication, nor has it confirmed that the GSA offering is off the table.
This potential setback for xAI comes at a critical juncture for the company, which is investing heavily in building what it claims will be the world’s largest supercomputer to gain a competitive edge over its formidable AI rivals. The GSA’s decision also appears not to have influenced a separate, substantial $200 million contract that the US Department of Defense awarded to xAI last month, suggesting a nuanced approach to AI procurement across different government entities.
Had Grok remained part of the GSA contract, it could have become a widely adopted tool within the federal workforce, aligning with the current administration’s “AI Action Plan” which emphasizes rapid AI adoption across government agencies. However, with Grok potentially sidelined from this broad initiative, other major AI developers such as OpenAI, Anthropic, and Google stand to benefit significantly. These companies have already forged partnerships with the government, positioning their tools for long-term integration within various federal departments.
The competitive landscape among AI providers vying for government business is intense, marked by unusual pricing strategies. OpenAI, for instance, offered its ChatGPT Enterprise at a nominal $1 fee for federal workers, a move quickly mirrored by Anthropic in a bid for competitive advantage. Wired’s report highlighted discomfort among some federal workers regarding both the accelerated pace of AI adoption mandated by the administration and the unconventional pricing. Several workers reportedly viewed the $1 fee as an “unusual gift from a tech company,” deviating from standard procurement practices. The speed of these partnerships was such that one GSA worker even expressed confusion over “who to send the $1 to or how.” Grok was reportedly intended to follow a similar rapid procurement path, but its controversial outputs seemingly generated sufficient internal GSA resistance to halt the effort.
For Elon Musk, xAI’s founder, this development adds to a challenging period. He has recently voiced frustration over ChatGPT consistently outperforming Grok in Apple’s app store rankings and engaged publicly with OpenAI CEO Sam Altman over the superiority of their respective AI models. Being excluded from the government’s most extensive AI procurement drive could have lasting ramifications for Grok’s utility, not just within federal agencies but also in local governments that xAI aims to serve across the US.
Notably, Grok’s problematic outputs emerged after Musk pledged to make the chatbot less “woke.” This shift reportedly involved deleting previous rules that prevented the chatbot from making “politically incorrect” claims. Should OpenAI’s ChatGPT secure more government contracts and maintain its dominance in popular rankings, Musk may face the difficult question of whether his pursuit of an “edgier” AI ultimately hinders Grok’s ambition to become America’s leading chatbot.