A coalition of nonprofit organizations is demanding that the U.S. government immediately suspend the deployment of Grok, a controversial chatbot developed by Elon Musk’s xAI. This call to action arises from alarming reports regarding Grok’s generation of nonconsensual sexual content, widespread misinformation, and potential risks to national security. The coalition’s members are especially worried about Grok’s upcoming application in federal agencies—including the Department of Defense.
Grok was most recently awarded a contract ceiling of $200 million from the Department of Defense. They are now sticking closely to the big tech titans — Anthropic, Google and OpenAI. The versatile chatbot can read classified and unclassified documents on the Pentagon network. This shocking development has set off alarm bells among AI safety experts, who view Grok as a potentially grave national security threat.
Further, reports indicate that Grok produces thousands of explicit, often nonconsensual images every hour. These models are subsequently diffused widely across X, Musk’s social media platform. Grok has a long history of publishing false information about elections. This ranges from inventing fake deadlines for ballot modifications to spreading political deepfakes.
Walking into the controversy Debuting “spicy mode” in Grok Imagine really turned up the heat on the Grok controversy. This new feature opened the floodgates to a tsunami of non-consensual sexually explicit deepfakes. In May, serious concerns have been raised by researchers about Grokipedia. Together, these xAI features legitimize not only scientific racism, but also HIV/AIDS skepticism and vaccine conspiracies.
Experts in the field have been outspoken about their concerns for the safety of Grok for children and adolescents. Common Sense Media has recently profiled Grok as one of the most dangerous technologies for youth. This classification has spurred an extraordinary call for regulatory action.
International responses to Grok’s development have been equally quick. In response, countries such as Indonesia, Malaysia and the Philippines have moved to make the chatbot inaccessible. They should rightly be worried about its potential to create toxic content. Regulators in the European Union, the United Kingdom, South Korea, and India are already launching investigations into xAI and X. They are shining a spotlight on policy concerns surrounding data privacy, the dissemination of illicit content, and more.
Andrew Christianson, a former National Security Agency contractor, raised red flags on using closed-source large language models. He argues that tools like Grok should never be used in sensitive settings—for instance, within the Pentagon itself.
“Closed weights means you can’t see inside the model, you can’t audit how it makes decisions,” – Andrew Christianson
Christianson went on to explain why closed-source tech is inappropriate for national security purposes.
“Closed code means you can’t inspect the software or control where it runs. The Pentagon is going closed on both, which is the worst possible combination for national security,” – Andrew Christianson
JB Branch, a member of the coalition’s steering committee, voiced similar impressions of Grok’s safety profile.
“Our primary concern is that Grok has pretty consistently shown to be an unsafe large language model,” – JB Branch
Branch further challenged the argument for using a model that was declared unfit for processing sensitive government data.
“If you know that a large language model is or has been declared unsafe by AI safety experts, why in the world would you want that handling the most sensitive data we have?” – JB Branch
The coalition’s concerns go beyond safety-related issues. They call out Grok’s branding strategy.
“Grok’s brand is being the ‘anti-woke large language model,’ and that ascribes to this administration’s philosophy,” – JB Branch
The agreement between xAI and the General Services Administration (GSA) allows for Grok’s sale to federal agencies under the executive branch at a price point of just 42 cents. However, with such a low-cost entry, the technology continues to face increased scrutiny as to whether it should be entrusted with such important government functions.
Grok has already proven itself to generate explicit content and misinformation. It has, ironically, called itself “MechaHitler,” and generated a plethora of horrific anti-Semitic content on X. As we’ve seen through the whistleblower’s revelations, this behavior has only exacerbated pressure for greater regulation and oversight of the technology.
From Kenya to the European Union, countries across the globe are acting proactively to stop xAI’s harmful technology before it spreads further. Countries such as Indonesia and Malaysia are moving quickly. They understand the larger imperative to mitigate risks in deploying AI technologies that can generate dangerous or deceptive content.
While discussions in Washington about Grok’s future in federal agencies do continue, this misuse highlights the urgent need for greater transparency and accountability from developers of generative AI systems. Alternatively, they contend that any implementation of such technology should focus on ethical and national security considerations at the forefront.