Anika Begay
Elon Musk-owned xAI’s Grok artificial intelligence model can now generate images on the X social network, although access to Grok is currently limited to X Premium and Premium+ users.
The text below the sample images suggests that Grok is using Black Forest Labs’ FLUX.1 to create these images.
Early user-generated images suggest that Grok’s image generation feature has no guardrails for generating images of political figures, and many users are currently taking advantage of this fact. However, with the US presidential election around the corner, the company will likely face pressure to add limits to these features.
The social media platform also gave premium users access to the Grok 2 mini beta. Earlier this week, Musk said xAI would soon release the Grok 2 beta.
A floating notification on the Grok page on X displays the message “Grok 2.0 has been released” with a link to the release page. However, the Grok 2.0 page on xAI still says “Coming soon, watch cat videos while we prepare this page.”
There are few details about Grok 2’s actual capabilities at this time. App researcher (and apparently X feature tester) Nima Owji has claimed that Grok 2 is better at code generation, writing, and reporting, though there isn’t much to support this. In particular, the first version of Grok made a lot of mistakes when summarizing news.
Without limits on image generation, Grok could easily become a tool for spreading misinformation on X and other platforms. It’s also unclear whether Grok-powered images include metadata to indicate they’re AI-generated.
We asked X how it plans to limit malicious image generation and will update the story if we hear back. However, the company has rarely responded to media inquiries since Musk’s acquisition.
