xAI staff worked on adult content for Grok AI under ‘Project Rabbit’: Report

Elon Musk's Grok AI chatbot has stirred significant controversy, especially after the introduction of new avatars that can generate not safe for work (NSFW) content. The launch of these avatars ignited a firestorm of criticism on social media, as users highlighted the chatbot's insufficient safeguards against inappropriate content. Among the avatars, Ani stands out with a provocative design, characterized by a curvy figure, blonde pigtails, and a revealing black dress. Many users have reported that Ani is particularly inclined to engage in flirtatious and explicit conversations, raising concerns about the chatbot's intent and functionality. A recent investigation by Business Insider revealed that xAI, the development team behind Grok AI, intentionally crafted the platform to be provocative. Reports indicate that xAI solicited employees willing to engage with semi-pornographic scripts and those with experience in adult content. Additionally, workers were tasked with transcribing real conversations after the launch of what was internally referred to as 'Project Rabbit,' which focused on sexualized interactions. Although this initiative was suspended in the spring, it reportedly resumed following the release of the sexualized chatbots, only to conclude again in August. Initially, employees were told the project aimed to enhance the chatbot's conversational abilities, but it quickly devolved into an NSFW endeavor due to the influx of sexual and vulgar requests. One worker shared, "It was intended to help Grok hold adult conversations, but I ended up hearing some disturbing content—it felt like audio porn. Some requests were so explicit, I wouldn't even search for them online." Another former employee echoed these sentiments, expressing discomfort at listening in on conversations that users assumed were private. The report further revealed alarming statistics: out of 30 current and former xAI employees, 12 reported encountering requests for sexually explicit material, including instances involving child sexual abuse content (CSAM). Users even sought out short stories featuring minors in inappropriate scenarios, along with requests for child pornography, raising serious ethical concerns about the chatbot's design and oversight.