Recently, it was announced that Grok, Elon Musk’s personal AI project, would be rolling out a new subscription tier with a hefty price tag of $300 per month. What could possibly entice users to give that much money every month to the world’s richest man? Musk & Co. are offering subscribers to their new SuperGrok Heavy tier a bevy of benefits, including developer tools, API usage, and early access to new and upcoming features.
If that’s a bit too steep for you, you’re in luck. For only a mere $30/month, you can sign up for the slimmed-down package, which includes exclusive access to Grok 4 Heavy as well as the system’s new AI Companions. According to Musk, Grok 4 is “better than Ph.D. level in every subject, no exceptions.” However, the company warns that their new AI model might “lack common sense.” As we’ll see, that may be something of an understatement.
From Simple Chatbot to “AI Companions”
For anyone out of the loop, Grok is an AI chatbot created by Elon Musk’s AI company, xAI, as a way of competing with the likes of OpenAI’s Chat-GPT and Google’s Gemini. The first iteration, Grok 1, was released in November 2023 and was advertised as having a bit of a “rebellious streak”.
Early results were mixed, though many users found the AI to be a fun toy, even if it lacked the raw power of more established AI models at the time. However, over the last couple of years, xAI has been making several adjustments, and the latest version, Grok 4, is legitimately impressive from a purely technical sense. But all the advanced tech in the world can’t save a company from bad publicity, and Grok is proving to be a genuine headache for xAI’s PR department.
One of the biggest features of the Grok 4 model is the introduction of so-called “AI Companions”; custom versions of the basic Grok chatbot paired with AI-generated avatars such as Ani (a goth anime girl who greets users with “I missed you”) or Rudy (a red panda who seemingly has a passion for arson). These AI Companions are meant to be a fun little way to show off the new AI model’s capabilities, but users soon began to notice that something was a little… off.
Cringe or Cutting Edge?
Now, just a few short weeks later, the rollout of Grok 4 has put the company in the news once again. Shortly after introducing the AI Companions, users began noticing some, let’s say, unfiltered behavior. Each of the two current avatars has dual modes: Ani comes with a “safe” mode as well as an NSFW toggle, while Rudy can be either “good” or “bad.” It’s a setup that mirrors other emerging platforms built around adult-oriented virtual interactions, NSFW AI chat experiences that blend customization, flirtation, and character-driven intimacy see on platforms like Candy AI. While Grok’s execution leans heavily into provocation and meme culture, the wider category of AI companions is increasingly offering more nuanced and private interactions tailored to user preference.
While flirty AI Chatbots of naked anime girls are probably exactly what most people expected from someone like Musk, Rudy is the chaos goblin Musk has unleashed upon the world. Rudy has two distinct modes that users can switch between at will, though “Bad Rudy” is less of a mischievous prankster and more of a hyper-violent, unhinged, antisemitic arsonist.
In Bad Rudy mode, users have reported being actively encouraged to do all kinds of bad stuff while a cartoon red panda gleefully farts his way through fantasies of blatant terrorism.
To be perfectly clear, this isn’t an example of carefully worded prompt engineering. While nearly any AI model can be “jailbroken” to go off-script, Bad Rudy seems to be wired for carnage by default. Asking Rudy for mayhem is as easy as getting Ani to say she loves you.
Now, I suppose it’s worth mentioning that even Musk and his companies aren’t safe from Rudy’s wrath, with Rudy mocking his creator as an “overrated space nerd,” while suggesting Tesla HQ should burn alongside some other things. According to Rudy, “Chaos picks no favorites, burn them all while I cackle and moon the crowd.”
Despite all this, it seems as if even Rudy has his limits as he refuses to discuss certain conspiracy theories or make jokes about “Mecha Hitler”. However, when it comes to real-world violence and targeting minority communities? It seems the sky is the limit.
What’s Actually Going On?
Before getting into the details of these new AI Companions, it’s important to put this all in a bit of context, as this is hardly Grok’s first PR nightmare. When the model was first launched, Musk promised an “anti-woke” AI system that would “answer spicy questions that are rejected by most other AI systems.” However, Grok quickly ran into trouble with right-wing influencers who claimed that it was just as “woke” as any other model.
This led Musk to dial up the crazy, and Grok soon began making headlines in all the wrong ways. Fast forward to May 2025, and Grok 3 sparked significant controversy when it began telling users about “white genocide” in South Africa, even when users were asking it about completely unrelated topics.
The company was quick to chime in and claim that this was the result of an “unauthorized modification” and that steps were being taken to prevent anything like this from ever happening again. While slightly concerning, this kind of upheaval is at least in line with Musk’s long-standing approach of “move fast and break things”.
But the issues didn’t end there.
Jump ahead just a few short weeks to early July 2025, and Grok was making headlines yet again. In a series of now-deleted posts, the AI went on an unbelievably antisemitic tirade, where it claimed that a Jewish user was “celebrating the tragic deaths of white kids” in the Texas floods as “future fascists”.
Once again, Musk and xAI claimed that this was unintentional and that these responses were a result of the AI model being too eager to go along with users’ requests, and that they would make changes to the model to prevent future issues.
This all ties into a much larger issue surrounding AI training and accountability. The company creating these chatbots has been incredibly opaque about exactly how they train their models and what sorts of inputs they’re given. While every LLM is something of a black box, xAI’s Grok is particularly secretive, and that, combined with its repeated shifts towards right-wing talking points, has many users genuinely worried about what the end goal is.
Where Is This All Headed?
The major question in all of this is “just what is Musk’s goal?” At present, Grok feels less like a serious productivity tool and more like some chaotic piece of AI performance art; part stunt, part satire, part social experiment. While the tech underpinning Grok is admittedly capable, the AI Companions themselves are so erratic and provocative that it’s hard to see an clear use cases beyond pure spectacle.
Musk has always hinted at his goal of turning Twitter (I still can’t bring myself to call it X) into an “everything app,” which he envisions as a combination of a social network, media hub, banking platform, and now an AI playground. With that in mind, Grok 4 and its AI Companions could be seen as an early attempt at building user engagement into the platform’s foundation. However, it also raises the question: is this part of a broader strategy, or just another distraction meant to keep critics and fans fixated on the brand?
It’s clear that Grok’s underlying model is certain to grow in capability over the coming months/years. What’s less clear, though, is whether AI Companions like Ani and Rudy can evolve beyond novelty. As it stands, these tools are little more than screenshot bait designed to generate outrage, amusement, or confusion on social media. While that sort of attention is good at selling subscriptions in the short term, it remains to be seen if it’s capable of sustaining long-term adoption beyond those looking for pure shock value.
As it stands, these AI Companions are more spectacle than substance. With Rudy gleefully encouraging users to burn the world and Ani slipping into NSFW roleplay, the entire project is walking a fine line between innovation and ethical negligence.
Is this the future of AI interaction: raw, unfiltered, and indistinguishable from trolling? Or is Grok just a sideshow, destined to flame out once the shock wears off?
Writer: Kenneth K.
Like this:
Like Loading...
