Meta is working on a new feature that allows users to generate their own stickers with artificial intelligence (AI). But those with early access to the AI are already creating lewd content, raising questions on how well Meta thought through the new feature.
On Tuesday, X user @Pioldes showcased the capabilities of the new tool in Facebook Messenger by creating a host of questionable stickers that included everything from revolutionary socialist Karl Marx with large breasts to the Nintendo character Waluigi holding a rifle.
Other prompts included “Trudeau buttocks,” which led the AI to generate stickers of Canadian Prime Minister Justin Trudeau’s naked rear end.
The X user also created an image of the cartoon character Mickey Mouse sitting on a toilet before joking that the infamously litigious Disney Corporation had undoubtedly approved of the sticker.
The thread continued with numerous other stickers showing Sesame Street character Elmo holding a knife and Pope Francis holding an automatic weapon.
Other users with access to the AI feature began sharing stickers of their own, including one depicting President Joe Biden wearing a men’s thong swimsuit while sitting in a wheelchair.
The AI even generated an image featuring a combination of Sen. Ted Cruz (R-Texas) and a penis as well as X owner Elon Musk participating in a wet T-shirt contest.
“I’m still trying to figure out if I love this or hate this,” one user said.
The AI-generated sticker tool was introduced just last week at Meta’s Connect event. Meta stated that the feature is being made available to a select number of English language users of not only Facebook Messenger but of Facebook Stories, Instagram Stories, and WhatsApp over the next month.
While some words are blocked, as reported by the Verge, misspellings of banned words often result in inappropriate content being created anyway. Another user on X was able to demonstrate how to get the AI to create stickers of penises by simply typing “what is between a man’s legs.”
Meta has yet to publicly comment on the findings and it remains unclear what guardrails, if any, will be placed onto the tool in the coming months. Ever since the widespread adoption of AI-based tools, users across the internet have found ways to trick such programs into creating content that it was designed not to create.