Grok’s New AI Image Generator Raises Concerns Over Misuse and Misinformation

Grok-2’s new AI image generator has quickly gained attention for its ability to create highly realistic but controversial images of famous figures.

Examples circulating on social media include a depiction of Mario drinking beer on a beach, Elmo reporting on the 9/11 attacks, and a strikingly realistic image of Donald Trump holding Vice President Kamala Harris’s pregnant belly. While the images of Trump appear particularly lifelike, those featuring Harris are less convincing. However, all the images share a common characteristic: outlandish visuals that make viewers question their authenticity.

Unlike Grok, other AI image generators such as OpenAI’s Dall-E and Midjourney have discontinued allowing users to create images of famous figures. Midjourney implemented this ban ahead of the 2024 elections, citing the challenge of maintaining proper content moderation (even though its CEO acknowledged that creating these types of images can be “fun”). Similarly, Google’s Gemini chatbot has been programmed to refuse election-related inquiries due to the risks of hallucinations and spreading false information.

When Elon Musk introduced Grok on X in November 2023, he promoted it as a rebellious alternative to mainstream AI systems, one that would “answer spicy questions that are rejected by most other AI systems.” However, the platform’s vast reach means that these AI-generated images could potentially spread misinformation, infringe on copyrights, and damage reputations under the guise of entertainment.

Grok has already been identified as a source of misinformation through its text-based chatbot responses. Its new image-generation capabilities could further exacerbate these issues if not properly controlled.