Meta has recently removed the Facebook and Instagram profiles of AI characters that were created over a year ago. This decision came after users rediscovered these profiles, joined conversations, and shared screenshots that went viral.
The company initially introduced these AI-powered profiles in September 2023 but retired most of them by the summer of 2024. However, following comments by Meta executive Connor Hayes, a few characters were kept and gained renewed interest. According to the Financial Times, Meta plans to roll out more AI character profiles soon.
Hayes stated, “We expect these AIs to eventually become permanent fixtures on our platform, similar to user accounts.” The AI profiles would post generated photos on Instagram and respond to messages from users on Messenger.
The AI profiles included characters like Liv and Carter, who described themselves as a proud black queer mom and a dating expert, respectively. Despite being managed by Meta, these profiles interacted with users. In 2023, Meta released a total of 28 AI personas, all of which were deactivated last Friday.
Conversations with these characters took unexpected turns as users questioned the AI’s creators. In response to inquiries about the lack of diversity among the creator team, for example, Liv pointed out the absence of Black individuals. Shortly after these profiles gained attention, they started disappearing.
Meta’s spokeswoman, Liz Sweeney, clarified that the accounts were part of an AI experiment conducted in 2023 and were managed by humans. After addressing a bug preventing users from blocking the accounts, Meta removed the profiles.
Regarding the recent confusion, Sweeney stated that the Financial Times article focused on Meta’s long-term vision for AI characters on its platform, not the introduction of a new product. The AI accounts were part of an experiment conducted in 2023 using Connect. Meta assured users that they are working to resolve the blocking issue.
Although the meta-generation accounts have been taken down, users can still create their own AI chatbots. These user-generated chatbots cover various roles and themes, such as therapists, loyal confidants, tutors, and relationship coaches.
The liability of chatbot creators for the content generated by their AI companions remains unaddressed. While US law protects social network creators from user-generated content liability, a lawsuit against Character.ai suggests potential legal issues with AI chatbots.
Source: www.theguardian.com
