Elon Musk’s Newest AI Invention Is a Goth Anime Girlfriend

It has been less than a week since Grok AI of Elon Musk did not declare himself “mechahitler” and made an anti -Semitic tirade in the responses of a user, but he already asks to be your girlfriend and tell the stories to your children. The Companions of AI was involved in legal and ethical debates roughly since their creation, but in a surprise article on X yesterday, Musk decided to throw his hat in the ring. Now, in the dedicated mobile application of Grok Ai, you can speak to “Ani”, a “Waifu” animates (their words, not mine) which looks in a shocking way to Misa Amane Death threatOr Rudi, a red panda “adapted to children” who will try to run everything you ask in a fanciful story.

The new “companions” are fueled by the vocal mode of Grok and are essentially animated characters displayed on specific vocal fashion personalities (a preexisting functionality). Musk’s Post indicates that they are intended for Supergrok subscribers, or people wishing to pay $ 30 per month for X AI, but when writing this article, I was able to test them despite being a free user. I left a little worried, but above all, I was just embarrassed.

Discuss with the animated companions of Grok

If you’ve already used Grok’s vocal mode, you will know what to expect here. You will get the usual responses and responses supplied by LLM usual filtered via the functionality of existing personalities of the objective of the vocal mode. For example, if you ask the “grok doc” personality on a strange taupe on your body, this will indicate potential risk factors, but if you ask for its “plot” personality, it will say that it is probably a government monitoring device. These differences in the responses are essentially only the invites to follow in the personality, and you can even create your own voice personalities only with personalized prompts at the moment. It is not as if Grok Doc knows more about medicine; He is just trained to act more professional.

Likewise, if you ask Ani on this same taupe, you will see an anime girl who thinks it’s a bit hot, but Rudi will recommend that you talk to an adult. Regarding the content, it is almost the same thing (the training libraries are not different), but now you have a character to look at, with a synchronization of the lips and a complete movement of the body. For my money, Rudy was the only personality with an appropriate response there – I still cannot trust AI with my physical or mental health, and the fact that Grok’s vocal mode even has predefined personalities who encourage the medical advice of AI or paranoid thought, even as a joke, is a sign that this application is probably not for me. It is however in expectations for Grok, and at least I received no factual error. The application has also lagged behind for me less frequently than the more generic Gemini of Google live, so technically, it seems solid.

At Grok’s credit, as much as a Bot AI with a “therapist” or “Sexy 18+” fashion sounds my alarm, most of my time with the new Elon companion robots was motivated by the same condescension, an overly pleasant conversation to people, almost all IA robots, tends to love. Ani threw more swear than what you might be used to, and Rudy could not stop talking about gommeux and rainbow, but little of the basic experience here has actually felt than new, and everything was even more “protocol right” than Her. You can get strange answers if you try, like when I became Rudy to make Bleets’ bar vapes, but the general experience over a few hours of use was quite banal.

I am also happy to announce that none of the companies AI has congratulated infamous dictators (a low bar, I know), or really stated a kind of controversial opinion in my time to speak to them, it therefore seems that there are guarantees in place, at least for the moment. In fact, according to the user of Bluesky Chadlog.md, you can really see the personalized prompts that have entered these companions yourself if you dig into the Grok code. Apart from Ani being explicitly “co -dependent” and by seeming suspiciously in Grimes (the guests teach that it is in alternative and independent music, and that it is “secretly a little a nerd, despite [her] An energetic appearance “), there are not many openly reprehensible things here. That said, I vomited when she said:” Imagine that you and I lying on a blanket, Dominus seated next to us “, in our first cat. (I learned later that Dominus is, apparently, a dog, something that I only discovered by reading the guests of the bot.)

Credit: X

Create in the parameters of the GROK application shows two more flip -flops linked to additional companions, one allowing “Bad Rudy” and one that allows NSFW content.

When I allowed the cursors myself, Rudy has essentially become a brand new companion, with a Southern park Style attraction for swear, coarse images and hatred for almost everyone – but still no Nazi praise. Ani, meanwhile, has been about the same, although the testingcatalog reports, you need to improve your romance with her to see NSFW content. This will be supposedly in her model of character with revealing lingerie, but for the time, she was always ready to engage roughly in sex with me. Oddly, she was also quite explicit in her conversations before going to the NSFW cursor, but it seems that there are bugs with this deployment, which Musk describes as a gentle launch. When I disabled the NSFW cursor, I did not have the opportunity to speak to Ani at all, so maybe I was not even supposed to see her before I light it.

Credit: X

A third chatbot AI, Chad, is also in preparation, and I can imagine what it would be just based on the name.

At worst, this is another step in the parasocial AI, with the same potential for catastrophic failure that we have seen before, the same problems as ordinary AI, and certain elements of manipulative gamification undoubtedly launched. Instead, even in its best, I always wonder what is the point of that.

I am someone who likes anime, but I cannot see a girlfriend of maniac robot which is impossible to eliminate the value of more than a few minutes of my time. It is essentially just Google without any credit to its sources and to a boring filter that I have to cross to reach everything that matters. And even then, I had to keep headphones so that my neighbors cannot hear it. Maybe I would have withdrawn more at the age of 14, but in my 1930s, it is an embarrassing way to spend my time which leaves me mainly cheesy. Targeting an audience of niche gooner is not the way I would reorganize all the Internet, but by going by Musk’s Rescosts, maybe I’m not big enough.

If you have a little more imagination than me, it might be more than new to you. But with the AI which always tries to settle like the next great thing, I will say that “cringing teeth” is not the first face that I would have put on it. These animated avatars are the most adjacent AI of the NFT which felt me in a certain time, and given how it is at the moment, it is not an excellent sign for the future of the industry.