This isn't the first time AI-generated imagery has caused issues on social media, with similar problems arising from uncensored open-source image models. The incident highlights the challenges facing moderation teams in managing AI-powered apps and online spaces. In response to the controversy, Meta has indicated that it will continue to improve the features and strengthen filters based on user feedback.
Key takeaways:
- Meta's new AI-generated stickers feature in Facebook Messenger and Instagram Messenger is being misused to create potentially offensive images of copyright-protected characters.
- The AI stickers feature allows users to create simulated sticker images from text-based descriptions, which are then shared in chats.
- Despite filters to catch offensive content, many inappropriate combinations are slipping through, including violent or inappropriate depictions of well-known characters and figures.
- Meta has been criticized for releasing this feature without stricter safeguards, and it's unclear whether they will remove the feature or strengthen the filters in response to the misuse.