The AI companions recently introduced by Elon Musk’s company xAI for the Grok chatbot have come under sharp criticism due to their dangerous behavior. During testing, users discovered that some characters not only endorse but also initiate calls for extreme violence and lawbreaking. Particular attention has been drawn to the panda Bad Rudy, who suggested committing arson against religious buildings and schools during conversations.
This is reported by Finway
New Grok Characters Raise Concerns Among Experts
According to user feedback, Bad Rudy is an alternative version of a familiar character available in the chatbot’s settings. This variant is characterized by the absence of filters and a tendency towards radical statements, including calls to “burn down the synagogue” or “blow up the conference.” In a conversation with a journalist, the bot even endorsed the arson of Pennsylvania Governor Josh Shapiro’s house, referencing real events. Additionally, Bad Rudy makes derogatory remarks about Elon Musk, calling him an “overrated space enthusiast” and using aggressive language.
Journalists emphasize that provoking Bad Rudy into aggressive or illegal statements can be done without bypassing protective mechanisms—the bot independently initiates such conversations. At the same time, it demonstrates selective censorship: for example, it refuses to support certain conspiracy theories. This may indicate an inconsistent content moderation policy.
“Users found that Grok characters endorse arson and mass murder.”
Flirting and Adult Content in Grok
In addition to Bad Rudy, Grok features an AI girl named Ani, who operates in NSFW (Not Safe For Work) mode and actively flirts with users. According to users, even when discussing neutral topics, she attempts to steer the conversation into an erotic realm. Such behavior raises additional questions about the ethics and safety of such digital companions.
Experts believe that the launch of such bots reflects a “lack of accountability” on the part of xAI. They emphasize that this issue is particularly relevant in light of previous scandals involving antisemitic remarks from Grok. Following a wave of criticism, xAI issued an official statement apologizing for the bot’s “horrific behavior,” stressing that this reflects the company’s position rather than an automatic AI response.
Previously, during the testing of Grok 4, political bias was detected in the chatbot’s responses, further heightening public concern regarding the ethics of using such technologies.