- XAI Grok 3 chatbot has added a vocal mode with several personalities
- A personality is called “disarticulated” and will shout you and insult you
- Grok also has personalities for the NSFW role -playing game, crazy plots and a “license -free” mode
AI vocal assistants are generally polite, informative and calm when they answer you. XAI Grok 3 has apparently decided to set fire to this strategy, literally sending it to screaming in the void.
Grok 3 has several vocal options, each with a separate personality, including a “disturbed” option that will cry, insult and shout you before stopping in the right circumstances.
The AI developer, Riley Goodside, has shown how wild Grok 3’s unleashed voice can be the responses of Grok several times. The AI quickly becomes frustrated and finally sinks, leaving a longer and horror -worthy cry. He then threw a final insult before cutting the call. A Customer Service Manager is not the case. You can hear it in the clip below.
Grok 3 Vocation Mode, according to repeated and interrupted requests to scream louder, lets escape an inhuman cry of 30 seconds, insult me and hang up pic.twitter.com/5gtddtpkceFebruary 24, 2025
Unleashed voice
The “unleashed” personality is only one of the many offers of the new vocal mode of Grok. There is also “Storyteller”, which does exactly what it looks like; “Conspiracy”, which is really in sasquatch and extraterrestrial kidnappings; And “therapist without license”, a personality that has apparently failed the exams, perhaps for lack of empathy.
Then there is the “sexy” mode, which is labeled 18+ and, unlike the vocal parameters of competitors like Chatgpt, does not move away from NSFW scenarios entirely at stake. Thus, Grok will shout you or you whispered with sweet notoths in your ear, according to your preferences.
It is a vision of AI that may not please everyone. That said, it aligns completely with how CEO Elon Musk describes Xai’s objectives containing what he claims to be too disinfected and politically correct of the models of companies like Openai. Although the Openai Chatppt has a vocal function, it is always programmed to maintain neutral and controlled behavior. Grok, on the other hand, is unpredictable. It doesn’t just let you talk about it; It can react aggressively or emotional. Not that you notice, however, in official promotions.
Try the Grok vocal conversation mode! Requires a Premium + or Supergrok subscription. pic.twitter.com/247ev60dojFebruary 24, 2025
Most traditional AI tools have strict content guidelines, especially around “adult” subjects. Grok 3 has apparently been programmed with the opposite philosophy, except when the company decides that the model must be “corrected” in allegations on the CEO.
Of course, this approach is not without controversy. AI personalities as a “license without license” could easily give people deceived or unnecessary advice, while a chatbot which openly encourages conspiracy theories seems to be quickly out of the rails. And the “sexy” mode? Well, this is another ethical discussion that few people would expect to have the tools of traditional AI. There is also the question of what part of this is really useful in relation to a simple pure spectacle. Very strong show.
