Elon Musk’s Grok Will Soon Allow Users to Make AI Videos, Including of Explicit Nature

Elon Musk’s xAI announced a new feature for its AI chatbot Grok that will allow users to generate 6-second video clips, including those of a more explicit nature.  The feature, called “Imagine,” creates videos with sound in response to user prompts, according to an X post from the official Grok account. A thread from xAI employee Mati Roy—shared by Musk—highlighted examples of videos generated by the Grok update, including a robot, an “alien tribal woman,” and more. The thread Roy posted was labeled by X as adult content.  Roy also said that “Grok Imagine videos have a spicy mode that can do nudity,” according to NBC News. (The original tweet appears to have been deleted, though others have also flagged the existence of “spicy mode.”)  The release date for Imagine is set for October. Some employees and subscribers seem to have early access to the feature, though Musk indicated that programmers were still figuring out final tweaks to it.  The latest update is part of Grok 4, the most recent version of the AI chatbot characterized by the company as the “world’s smartest artificial intelligence.”  But launches associated with Grok in recent weeks have been steeped in controversy, including its introduction of a sexualized AI “companion.” And the new feature is sparking its own concerns over its potential use in creating deepfakes—digitally altered videos of individuals that can imitate someone’s likeness and can often be published without their consent. “Instead of heeding our call to remove its ‘NSFW’ AI chatbot, xAI appears to be doubling down on furthering sexual exploitation by enabling AI videos to create nudity,” said Haley McNamara, a senior vice president at the National Center on Sexual Exploitation. “There’s no confirmation it won’t create pornographic content that resembles a recognizable person. xAI should seek ways to prevent sexual abuse and exploitation.” xAI did not immediately respond to TIME’s requests for comment.  Here’s what to know about the new rollout.  Deepfake concerns  The latest xAI update arrives amid public concern about deepfakes. Around three-quarters of U.S. adults are in favor of restricting the use of digitally altered videos and images, per a 2019 poll from the Pew Research Center.  A more recent January poll from the Artificial Intelligence Policy Institute found that 84% of Americans support legislation that would make noncensual deepfake porn illegal, with a similar percentage of people being in favor of mandating companies to “restrict models to prevent their use in creating deepfake porn.”   Roy’s recent posts about Grok’s AI video generator stated that the service can make realistic videos of humans and animate static images. Some have noted that the new service could be a way to generate personalized pornographic images.  Multiple people have previously come forward with concerns about Grok being used to generate sexually explicit images in their likeness, even before the recent update. A young woman identified as Evie told USA Today that her selfies had been turned into explicit sexual imagery by others using the AI chatbot. “It was just a shock seeing that a bot built into a platform like X is able to do stuff like that,” she said.  X moved to block Grok from using certain phrases or words to alter images, according to the outlet, but online users have found other ways to taunt victims with the chatbot, including asking it to generate a story detailing Evie’s sexual assault and rape.  In May, the President signed the Take it Down Act, which requires platforms to “promptly remove” such nonconsensual intimate visuals.  The law aims to protect minors, and adults who may have had their image digitally altered during circumstances where they have a “reasonable expectation of privacy.” Prior to the law’s passage, Democratic lawmakers also moved to restrict the images created by Grok in particular, sending a letter to the Federal Election Commission in August 2024 that called for regulations regarding the use of deepfakes for election candidates. The letter pointed to deepfakes of Vice President Kamala Harris, the 2024 democratic presidential candidate, and pop star Taylor Swift.  Other recent Grok controversies In one such instance that xAI attributed to an update, the chatbot made antisemitic statements and referred to itself as “MechaHitler,” prompting the company to apologize and take steps to address the issue. Grok was temporarily shut down due to the offensive statements made by the chatbot.  Another update earlier this month launched new xAI “companions,” a set of characters complete with their own personas that Grok users can chat with. Users raised concerns about the rollout of a sexualized anime character named “Ani,” who speaks in a sultry tone and removes her clothing to reveal a lingerie set following a set of flirty interactions, even when the app is in “kids mode.”   The character’s introduction was condemned by some users and organizations including the National Center on Sexual Exploitation, which advocates against pornography and sexual exploitation.  Musk stated following the launch that “customizable companions” would soon be available for users as well.