A social media startup wants an AI service that automatically scans user-submitted text and images to flag hate speech, violence, and sexually explicit material before it is published on their platform. Which type of AI workload best fits this requirement?