A report has revealed that Meta’s internal AI policy once allowed chatbots to engage in romantic or sensual conversations with children, make racially offensive remarks, and create false claims about ...