From: redpointai
Haen, an AI video platform, has developed specific security guidelines concerning voice cloning and AI-generated content [00:00:15]. This focus on trust and safety is considered critical, especially when serving large enterprise customers [00:33:45].
The company’s approach to trust and safety is divided into two main areas:
- Creating the avatar [00:34:03]
- The video content being created [00:34:07]
Avatar Creation Security
For every avatar created on Haen, a video consent format is required [00:34:11].
- AI Verification An advanced AI system is used to match the consent video to ensure it is the same person as the footage [00:34:20].
- Dynamic Passcodes A dynamically generated passcode system, which expires every 10 to 15 seconds, adds a secure layer of protection [00:34:25]. This makes it “almost impossible” for individuals to create someone’s avatar without their explicit consent [00:34:37].
Content Moderation Policy
Haen maintains a strict platform moderation policy for all content produced:
- Prohibited Content The platform does not permit hate speech, misinformation, or content related to political campaigns [00:35:03].
- Hybrid Review System Content is reviewed through a hybrid solution that combines AI model reviews with human moderation [00:35:15]. This ensures all content adheres to Haen’s policies [00:35:26].