Grok-Chats-LeakedPrompts-on-Drug-Manufacturing-&-Killing-Musk-Viral

Wild Grok Prompts on Making Drugs and Killing Elon Musk Went Viral with xAI’s New Sharing Feature

Elon Musk’s Grok AI is under scrutiny as several thousand Grok AI chats went public on Google search. These prompts involved everything from questions on mundane tasks to performing illegal activities and plotting murder.

A report published by Forbes mentioned that the users who clicked the “share” button ended up permitting search engines to access their chats and that every conversation shared had a unique URL visible to Google, Bing, and other search engines.

How Did Grok AI Chats Become Public?

The share button available on the Grok interface made all the chats public and accessible by search engines without properly informing the user. This made more than 370,000 Grok prompts vulnerable to indexing.

A similar incident occurred with ChatGPT, where the ‘share’ button was enabled by default, allowing the AI tool to share conversations with search engines. Many users unintentionally made their private and sensitive data public. 

However, OpenAI responded to the users' criticism by saying that it was just a short-lived experiment and that the share button would soon be removed. Elon Musk, who was quick to take a jibe at OpenAI for such carelessness, ended up facing the same reproval. 

What Prompts Were Given to Grok?

While AI tools are actively being used for writing code, creating content, and drafting emails, the leaked Grok AI chats have shown that users have taken the bot’s utility up a notch. The chatbot has now turned into a patient listener, a medical consultant, and much more.  

People have risked sharing their private medical and psychological reports, causing ethical concerns. The prompts also included some horrifying requests like planning suicide, assaulting Musk, developing malware, and promoting racist content. 

Though xAI shouldn’t have made the user prompts public without their permission, the nature of the leaked Grok prompts, especially the detrimental ones, has provided insight into the extent to which people have begun exploiting the AI.

To sum up, a “share” button in Grok AI allowed users to share their chats with search engines without any disclaimer. Google and other search engines indexed more than 370,000 conversations. These leaked prompts gave a glimpse at how people have started using AI and raised serious concerns.