People are figuring out ways to generate inappropriate images with Meta's new AI sticker feature
Last week, Meta launched its new AI sticker feature. Users have found ways to generate wildly inappropriate images with the tool, and are sharing the results online. One academic hit out at Meta's AI safeguards in a blog post on Sunday.
Controversy Surrounding Meta's AI Stickers
Meta launched the AI stickers and other features like celebrity AI assistants on September 27. The company stated that this new feature allows users to turn "text prompts into multiple unique, high-quality stickers in seconds" on Meta apps like Instagram and Messenger.
However, the stickers have already sparked controversy online as users have been trying to generate inappropriate images using the technology. Some examples shared include a child bearing a gun and a lewd image of Canadian Prime Minister Justin Trudeau.
Response on Social Media
User @pioldes shared examples of the imagery created with the new AI tool, which quickly gained attention on social media platforms. Subsequently, other users began sharing their own stickers online, including images featuring popular characters such as Mickey Mouse holding a gun, Elmo smiling with a knife, and even a pregnant Sonic the Hedgehog.
Despite certain words being blocked by the feature, users have found ways to bypass these blocks using typos or creative prompts.
Criticism from Academia
Tama Leaver, a professor of internet studies at Curtin University, expressed criticism of Meta's AI safeguards in a separate blog post. Leaver questioned the effectiveness of Meta's safeguards, highlighting instances where the AI produced inappropriate content despite supposed restrictions.
The AI-generated stickers are powered by Llama2, Meta's competitor to ChatGPT, and were initially rolled out to select English-language users with gradually increasing safeguards.
Concerns Beyond Meta
This issue of problematic AI-generated content is not unique to Meta. In the past, the AI avatar generator app Lensa faced allegations of sexualizing and racializing user avatars, despite rules against adult content.
Representatives for Meta, Disney, Nintendo, Sega, and Trudeau have yet to respond to requests for comments regarding these issues.
Read the original article on Business Insider
