Introducing Futuresafe, our new responsible creation report

Today, we’re publishing the first edition of our yearly transparency report, providing insights into how we ensure our platform is being used responsibly, in compliance with existing online safety regulations and our own policies.

For the past five years, we’ve empowered tens of thousands of businesses to harness the power of AI video technology safely, responsibly, and with unmatched integrity. This commitment is especially critical as many of our customers are Fortune 1000 companies operating in regulated sectors such as healthcare, finance, and automotive industries — fields where trust, compliance, and accountability are paramount.

Synthesia stands apart as the only generative AI video platform that implements rigorous content moderation at the point of creation and operates based on an ISO 42001-certified AI management system and on industry-leading AI governance practices. 

Our moderation system is designed to offer multi-layered and around-the-clock protection, by combining sophisticated automation with human oversight. In 2024 alone, the automated components of our moderation system parsed through approximately 6.5 million videos, resulting in the proactive removal of over 314,000 pieces of content.

Supplementing this, our dedicated moderation team, providing continuous 24/7 coverage, reviewed nearly 793,000 flagged items, rejecting approximately 139,000 that violated our robust Acceptable Use Policy (AUP) and content policies.

Recognizing the sensitivity around specific content types, Synthesia applies context-specific restrictions, particularly around political or news-like content. Such content is limited to enterprise users employing custom avatars that go through extensive security checks before getting access to the platform. This approach is designed to ensure that even seemingly benign activities, such as referencing current events in a video about workplace safety (e.g. “There are news reports of a weather event affecting our warehouse, please be careful when travelling into work”), undergo careful review to maintain platform integrity and compliance.‎

We deeply value transparency and user empowerment. As part of our extensive moderation process, users receive clear reasons when content is restricted and have the ability to appeal decisions. In 2024, we processed over 5,500 moderation appeals, approving 21.5% of them. Appeals are handled efficiently, with responses delivered within 48 hours, reinforcing our commitment to fairness and transparency.

Our approach has been rigorously validated externally through comprehensive tests and reviews, including by the Responsible Innovation Labs, Risky Business Solutions, and through the first-ever public red-team test conducted by the National Institute of Standards and Technology (NIST) in the United States, in partnership with Humane Intelligence. These assessments affirmed the strength and effectiveness of our moderation and security practices.

Synthesia continues to lead with an unwavering commitment to responsible AI by proactively moderating content, continuously enhancing its safeguards, and striving to keep its platform safe and secure.

You can access our 2024 transparency report from our Trust Center.