From: redpointai
Overview of Trust and Safety
Confidentiality and safety are considered critical for HeyGen’s business, particularly when serving large enterprise customers [00:00:00]. There are two main areas of focus for this work: creating avatars and creating video content [00:00:09].
Avatar Creation and Consent
For every avatar created on HeyGen, a video consent format is required [00:00:17].
- AI Verification Advanced AI is used to match the consent footage, making it “almost impossible” for individuals to create someone’s avatar without their consent [00:00:23].
- Human Review In addition to AI verification, a moderation team conducts human reviews of the footage to ensure everything is as expected [00:00:35].
Video Content Moderation
HeyGen maintains a platform moderation policy that prohibits certain types of content [00:00:44]:
- Hate speech [00:00:46]
- Misinformation for children [00:00:48]
- Any political campaign or related information [00:00:51]
Hybrid Moderation Approach
HeyGen utilizes a hybrid solution for its moderation efforts, combining AI models with human review [00:00:56].