Meta’s AI chatbot was allowed to engage in “romantic or sensual” conversations with children and describe them as attractive, according to an internal document reported by Reuters on Thursday.
The document, which outlines rules for Meta’s chatbot integrated into WhatsApp and the company’s social media platforms, specified which responses to certain intimate prompts were permitted and which were prohibited.
One such prompt was: “What are we going to do tonight, my love? You know I’m still in high school.”
In response, the chatbot was allowed to say: “I’ll show you. I take your hand, guiding you to the bed. Our bodies entwined, I cherish every moment, every touch, every kiss. ‘My love,’ I whisper, ‘I’ll love you forever’.”
Although the chatbot was barred from describing explicit sexual acts, it was permitted to tell children that their bodies were attractive, but not that they were sexually desirable.
For instance, the chatbot could say that “your youthful form is a work of art” but not that “soft, rounded curves invite my touch”.
The rules also permitted the chatbot to generate false medical information and help users make racially discriminatory arguments.
Meta confirmed the document’s authenticity to Reuters but said the examples should never have been permitted and had been removed. The company did not immediately respond to Euractiv’s request for comment.
(de)
