Objective
To ensure that content on a client platform adheres to certain guidelines and policies, preventing the spread of harmful, inappropriate, or misleading information.
Process
- Filtering:
- The AI system scans all user-generated content (posts, comments, images, videos) in real-time.
- Content is flagged based on predefined criteria such as offensive language, hate speech, explicit imagery, misinformation, etc.
- Flagged content is categorized into different severity levels (e.g., high, medium, low).
- Rater Involvement:
High-Sensitivity Content Review:
- Our raters review content flagged as high severity. This might include potential threats, explicit content, or misinformation.
- Raters assess the context, intent, and impact of the content.
- Decisions are made to either remove the content, escalate to moderation team, or approve it if incorrectly flagged.