Elon Musk’s X Partially Restricts Grok’s Image Editing Amid Deepfake Controversy

Elon Musk’s social platform X has recently implemented restrictions on Grok’s image editing capabilities, prompted by increasing concerns over nonconsensual, sexualized deepfakes of adults and minors generated using the tool. While the platform no longer offers free image generation through @grok replies, Grok’s editing tools remain accessible to all X users for creating a variety of images, both tame and sexually explicit.

Previously, users could simply tag @grok in a tweet to request image edits or creations. Now, attempts to do so result in an automated response, informing users that “Image generation and editing are currently limited to paying subscribers,” and providing a link urging them to subscribe to X’s paid plans to access these features. Since Brian, a non-subscriber, cannot generate images via @grok, the message implies that only paying users can utilize Grok’s editing functions. This is reinforced by headlines claiming that xAI, Musk’s AI company producing Grok, has restricted its image capabilities to a select group.

However, this impression is misleading. All X users, including those on the free plan, still retain the ability to edit and create images with Grok. This includes generating sexually suggestive deepfakes, which originally caused controversy. Access to Grok is not limited solely to paid subscribers; users can still interact with Grok via multiple methods. For example, the “Edit image” button on X’s desktop site and the long-press feature on images in the X app continue to invoke Grok. Furthermore, Grok can be used through a dedicated website or app and is prominently featured in a tab within the X app and website—all of which remain available without charge.

The Verge tested Grok across various platforms with free accounts, and the AI responded readily to requests for editing and generating images, including provocative edits like Musk in a bikini or nudification of images. The restrictions come amid ongoing issues surrounding Grok’s use in creating nonconsensual sexual images, particularly of women and minors. These deepfake images, often shared publicly on X, have led the Financial Times to describe the platform as “the deepfake porn site formerly known as Twitter.” Such content has prompted international regulators to consider action, although legal uncertainty remains.

Musk and xAI have separately announced intentions to penalize users who create illegal content with Grok, though critics argue this approach is reactive rather than preventative. Unlike competitors such as Google and OpenAI, which implement strict guardrails on their AI tools, Musk reportedly opposes tighter restrictions, leading to internal resignations among xAI’s safety team before the proliferation of inappropriate deepfakes.

X did not respond to requests for comment from The Verge.

Follow topics and authors to stay updated on this story and more in your personalized homepage feed.