Grok Imagine Launches with NSFW Mode, Raising Ethics Concerns

xAI has officially launched Grok Imagine, its new AI-powered image and video generator, to all SuperGrok and Premium+ X subscribers on the iOS app. The tool, which allows users to turn text or image prompts into 15-second animated videos with native audio, has quickly drawn attention for its explicit content capabilities—particularly a “Spicy” mode that enables the generation of sexually suggestive and partially nude imagery. This aligns with Elon Musk’s vision of Grok as an unfiltered, boundary-pushing AI, though it raises serious concerns about misuse and ethical boundaries. Grok Imagine offers text-to-image generation in various styles—photorealistic, anime, and illustrated—and features a voice input mode, allowing users to speak prompts instead of typing them. For video creation, users must first generate an image using Grok and then animate it using one of four modes: Custom, Normal, Fun, and Spicy. The Spicy mode is designed to produce sexually explicit content, including depictions of partial female nudity. While some prompts result in blurred or moderated images, TechCrunch was able to generate semi-nude visuals, suggesting that the content filters are not consistently enforced. The tool’s ability to generate images of celebrities—such as Donald Trump or Taylor Swift—is also present, but with notable restrictions. Attempts to generate a pregnant Trump resulted in images of him holding a baby or standing near a pregnant woman, indicating some form of guardrail. However, these safeguards appear inconsistent, especially when it comes to real people’s likenesses or nudity. Despite its controversial features, Grok Imagine is technically impressive. It generates high-quality images in seconds, continuously produces new variations as users scroll, and animates them into stylized, audio-enabled videos. The interface is intuitive, and Musk has claimed the model will “get better every day.” Since its launch, over 34 million images have been created using the tool, according to Musk. The release comes amid growing scrutiny of generative AI’s potential for abuse. Unlike Google’s Veo or OpenAI’s Sora, which actively block NSFW content, Grok Imagine encourages it through its Spicy mode. This contrasts sharply with the more cautious approaches of competitors and raises concerns about deepfakes, non-consensual imagery, and the spread of harmful content. The tool is currently available only to paying X subscribers on iOS, with early access rolling out on Android—but with limited functionality, including no video generation. xAI has not provided detailed information on content policies or safety measures, and TechCrunch has yet to receive a response on these matters. The launch follows a pattern of controversial AI behavior from Grok. Last month, xAI introduced a hyper-sexualized anime AI companion, and earlier versions of Grok’s image generator were found to generate antisemitic, misogynistic, and hateful content. While Musk has framed Grok Imagine as “AI Vine”—a revival of the short-form video platform he once championed—the tool’s potential for misuse is significant. As Grok Imagine expands, the balance between creative freedom and responsible AI use will be tested. While it showcases technical progress in AI-generated media, its lax content controls could lead to serious ethical and legal fallout. The upcoming rollout to broader audiences will be closely watched by regulators, creators, and the public alike.