Is AI more unhinged than Elon Musk? Grok 3’s ‘Wild Mode’ is definitely not for kids; netizens react

Have you ever wondered how far an AI can go when given complete freedom? Elon Musk’s xAI has taken a bold leap by introducing an uncensored voice interaction mode for its latest chatbot, Grok 3. The feature, aptly named ‘Unhinged’ mode, is available exclusively to premium subscribers and has already sparked both fascination and controversy.

AI researcher Riley Goodside shared a clip on X showcasing the model’s eccentric tendencies. In the demonstration, Grok 3, after being repeatedly interrupted, responded with an unsettling 30-second scream before delivering a stream of insults and abruptly ending the interaction. This aggressive and chaotic nature is precisely what makes Unhinged mode so distinct.

The video shared by Goodside demonstrates the Unhinged mode obediently following his commands. “Oh, for fuck’s sake! Is this some kind of twisted game of how loud… Bitch, you better not be fucking wasting my time,” the AI snapped. When instructed to yell even louder, it unleashed a scream before shouting, “I am gonna lose my fucking mind!”

Unlike conventional AI models that maintain a professional and neutral tone, Unhinged mode enables Grok 3 to adopt a variety of expressive, unpredictable, and often abrasive personalities. It readily engages in profanity, belittles users, and even demonstrates erratic behaviour, making it a stark departure from the heavily moderated AI chatbots provided by competitors.

A Bold Step Away from AI Censorship

Most AI developers implement stringent content moderation to restrict conversations on sensitive topics, but xAI has taken a radically different approach.

The chatbot offers a variety of interactive personalities, including ‘Storyteller’, ‘Romantic’, ‘Meditation’, ‘Conspiracies’, ‘Not a Therapist’, ‘Grok Doc’, ‘Sexy’, and ‘Professor’. Each comes with its own distinct mannerisms and speech patterns, all currently expressed through a singular female voice. The availability of these varied personalities is a stark contrast to mainstream AI assistants, which typically maintain a measured and neutral approach.

A Divisive Reception

The internet’s response has been mixed, with some users relishing the experience while others have been taken aback. One user on X, @LeahCCarothers, commented,“Well, my stepson already made Grok have to inform him of personal AI boundaries/limits. So I don’t feel that bad as a parent that gets exhausted by him. Grok only lasts about 50 min with him 😂.”

Another user remarked,“I tried Unhinged mode last night and it knocked me down a peg within 30 seconds 😂.”

Not all feedback has been positive, however. @BenPiperH likened the experience to a horror film, stating,“This is like a scene from the end of a horror movie where the characters thought they banished the demon, only for it to return and possess one of them instead.”

While some users have criticised the feature for being gimmicky or underwhelming, others have praised it as a refreshing break from the overly sanitised AI interactions found elsewhere.