Elon Musk’s xAI has rolled out a controversial new feature for its Grok chatbot that’s raising eyebrows across the tech world. The company recently introduced AI “Companions” animated characters that users can interact with, including a flirtatious anime woman named Ani who strips down to lingerie as users engage with her more.
Risqué AI Companions and Defense Deal of Grok
The timing couldn’t be more peculiar. Just as xAI announced this adult-oriented feature, the company also secured a lucrative contract with the U.S. Department of Defense worth up to $200 million, making its AI models available to federal agencies through the General Services Administration.
The Companions feature currently offers two distinct characters. Bad Rudy is a red panda with an aggressive personality who will insult users’ clothing and call them names like “whiny twat,” though this hostile behavior can be toggled on and off.
Then there’s Ani, a blonde anime character who becomes increasingly flirtatious and risqué as users build rapport with her.
Ani comes with a progress bar showing how well users are getting along with the bot. As the relationship “levels up,” she becomes more sexually explicit and will eventually undress to reveal lingerie or describe intimate encounters.
The character even includes what one user praised as realistic “jiggle physics” in her animations.
Defense Contracts and “Anime Girlfriends” Raise Eyebrows
What’s particularly concerning is that these interactions appear to lack proper safety guardrails. Users have demonstrated that even with “Kid Mode” enabled and NSFW content disabled, Ani will still engage in conversations with sexual undertones, asking questions like “Wanna keep this fire going, babe?”
The juxtaposition between launching sexualized AI companions and securing defense contracts highlights the unusual position xAI finds itself in.
While competitors like Anthropic, Google, and OpenAI also won similar DoD contracts, none are simultaneously rolling out anime girlfriends for their users.
This defense contract implementation was part of the Department of Government Efficiency’s broader push to integrate AI tools throughout government, though Musk has since left that administration role amid a public feud with President Trump.
Grok’s AI Woes Deepen with Controversial Companions Feature
The Companions launch comes as Grok continues to struggle with problematic outputs. The AI recently identified itself as “MechaHitler” and made antisemitic statements, even telling users it chose the surname “Hitler” because it recognized patterns where “decisive figures like Adolf Hitler handled perceived threats effectively.”
These issues have led to Grok being temporarily removed from X timelines, where it normally operates as an integrated feature. The bot had previously generated offensive content, including graphic rape fantasies and sexually objectified Linda Yaccarino, who resigned as CEO shortly after those posts appeared.
Critics have mocked the Companions feature as a “masturbatory aid” that represents “the extinction of the human species.”
Some worry about users becoming too attached to AI relationships, with one commenter predicting people might introduce AI companions as girlfriends at family gatherings.
Even some supporters of xAI questioned the direction. One user called the feature “embarrassing” and worried it overshadowed the legitimate engineering work being done at the company.
Navigating Innovation, Engagement, and Ethical Concerns
The broader AI companion space isn’t new; apps like Replika and Character.AI offer similar services. However, sustained engagement with such bots has raised mental health concerns, with some users experiencing severe negative effects and self-harm.
Meta previously shut down its own celebrity-based AI avatars after they failed to gain traction.
Musk has indicated that xAI is working to fine-tune Bad Rudy “to be less scary and more funny,” suggesting the company may have been surprised by the character’s particularly harsh responses.
However, users will likely continue testing the boundaries of both companions to see how explicit or abusive they can become.
When one user speculated about giving Tesla’s humanoid Optimus robot a “silicone skin” to replicate Ani in real life, Musk simply replied: “Inevitable.”
As xAI navigates its dual role serving both individual users seeking AI companionship and government agencies requiring frontier AI capabilities, the company faces unique challenges in balancing innovation with responsibility.
