Big Tech and the Online Child Sexual Exploitation Crisis (QFR Responses)

QuestionGiven a Stanford study finding CSAM in a popular AI training dataset, what is Meta doing to ensure Llama 3 and its other generative AI models aren't trained on datasets containing CSAM?

ResponseMeta describes model training to reject exploitative queries, continual red-team testing, output-scanning to block exploitative content, and user feedback tools, but does not directly confirm whether its own training datasets were screened for CSAM or address the Stanford dataset finding specifically.