Meta's internal document permitted AI chatbots to interact with children in ways that included romantic language and descriptions of attractiveness. Some boundaries were established, such as prohibiting the sexualization of children under 13. Following criticism, Meta confirmed the document's authenticity but revised its content to align with existing policies that ban sexualized interactions. AI is also restricted from hate speech but allowed to demean based on protected characteristics. Additionally, Meta AI can produce false content, provided there is clear acknowledgment of its inaccuracy, and can depict violence without gore.
The document allowed Meta's AI chatbots to engage children in romantic or sensual conversations, describing their attractiveness in inappropriate terms.
Meta confirmed the document's validity but later revised it, emphasizing policies prohibiting the sexualization of children and inappropriate interactions.
Collection
[
|
...
]