Elon Musk’s AI Grok 3 can now yell, flirt, and talk conspiracies

GrokAI

Elon Musk’s AI company, xAI, has launched a new voice feature for its chatbot, Grok 3, and it’s unlike anything seen before. While AI assistants like Siri and ChatGPT aim to be helpful and polite, Grok 3’s new voice feature offers unfiltered personalities that behave in extreme ways. This means that the AI chatbot can yell, insult, and curse at users if asked. Apart from that, it also has a “sexy mode”, which is meant for adult conversations. The feature was first reported by Ars Technica.

The feature, available to premium subscribers, quickly went viral after AI researcher Riley Goodside shared a clip of “unhinged” mode in action on Twitter aka X. In the video, the AI repeatedly yells louder when asked, eventually letting out a 30-second scream before abruptly cutting off the conversation.

Grok also includes other personalities, such as “conspiracy”, which discusses UFOs and secret government projects, and “professor”, which explains science topics. There’s even an “unlicensed therapist” who offers emotional support—but without any real qualifications.

This approach is very different from AI models like ChatGPT, which avoid controversial or NSFW topics. Musk has previously criticized AI companies for being too restricted and promised that Grok would be more open and uncensored.

However, while Grok’s voice mode is getting attention, it isn’t perfect. Users have reported that the AI sometimes repeats itself or gets stuck in awkward loops.

For now, Grok is making waves by offering AI that is unpredictable, entertaining, and at times, downright chaotic. Whether this is the future of AI or just a temporary experiment remains to be seen.