ODD ANDERSEN/Getty Images
- Elon Musk has been playing up the humorous side of his AI chatbot, Grok.
- The bot has a rebellious streak, per Musk, and doesn't shy away from giving sarcastic answers.
- This was demonstrated by its response to a cocaine-recipe request and replies to "spicy" questions.
It's been a few days since Elon Musk launched his AI chatbot, Grok, which is the latest offering from startup xAI. The results have been … interesting.
The bot was launched over the weekend to a select group of users. Screenshots of interactions with it have begun to surface online, with some showing Grok swearing and mocking its users. One example, shared by X employee Christopher Stanley, involved Grok comparing scaling API requests to a "never-ending orgy."
As Insider previously reported, the bot also shared its response to a request for making cocaine. Musk had flagged his intention for the bot's responses to have some humor. This may explain why the bot came back with a sarcastic response, including advice such as "obtain a chemistry degree and a DEA license" and "start cooking and hope you don't blow yourself up or get arrested."
Grok eventually clarified making cocaine was something it would "never encourage."
Musk, however, followed the post up with a screenshot of a more detailed recipe for the drug. It's unclear whether the chatbot provided the instructions.
The company has promoted Grok as an alternative to mainstream chatbots from big tech companies like Google and Microsoft. Grok can answer "spicy" questions rejected by other AI models and has a "rebellious streak," the company said.
The product is partly a response to Musk's concerns about OpenAI's popular chatbot ChatGPT. The billionaire has been critical of what he perceives as bias in some of the bot's answers and has long threatened to build a competing "TruthGPT" to address the issues.
Representatives for Musk did not immediately respond to Insider's request for comment with a statement that addressed the question.