Our AI Governance Framework
Learn how Synthesia addresses legal, procurement, and security requirements with built-in safeguards and clear governance practices.
Trusted by over 70% of Fortune 100 companies.
Responsible AI is part of our DNA
Grounded in the principles of Review, Report, and React, our 3Rs framework guides how we build and govern our platform.
Review
Governance starts long before a product launches. For us, it is a continuous process embedded in every stage of our platform, from risk assessment before a product decision to the audit trail that follows deployment.
Risk is not uniform. That’s why we apply proportionate controls by use case, distinguishing routine internal content from higher-risk scenarios like impersonation or manipulation.
Report
We are building a platform where content provenance strengthens enforcement and ensures accountability at every level.
We also maintain shared ownership of AI security across policy, engineering and customer success. Externally, we work with model providers to integrate their products safely, and keep customers informed about our technology choices
React
We built the platform to be used by thousands of organizations — and we designed it with security in mind. That means catching issues early and responding quickly.
Since day one, we made the decision to moderate content at the point of creation, not clean up after distribution. Regular audits, incident readiness, and a commitment to getting better ensure our governance keeps pace with the technology.
Turning Principles Into Practice: Synthesia’s Governance in Action
Trust & Assurance
- Terms of Service - terms that are clear, balanced, and tailored to our service
- Data Processing Agreement - the most stringent privacy laws for processing data
- Enterprise-Grade Security - robust, certified security practices for peace of mind
- ISO 42001 Certification - international standard for responsible development and use of AI
- Acceptable Use Policy - rules that govern the responsible use of our service
- AI Governance Practices - transparent governance practices we abide by
- Transfer Impact Assessment - standards for safe and secure cross-border data transfers
- AI Impact Assessment - thoughtful AI development in collaboration with diverse stakeholders
- Data Protection Impact Assessment - rigorous assessment of privacy risks in handling personal likeness
- Environmental impact - AI video as a carbon-efficient alternative to traditional video production
Content & Delivery
- Content Moderation - systems in place for proactive prevention over reactive remediation
- Avatar consent - consent safeguards for creating custom avatars and voices
- Avatar sharing - permission-based sharing of custom avatars and voices
- Single Sign-on (SSO) - streamlined login with enhanced account security
- Brand Kits - consistent and on-brand visual identity for your videos
- Workspace Settings - customizable permissions across distinct workspaces
- Live Collaboration - real-time collaboration with multiple stakeholders
- Accessibility - inclusive design through colors, captions, audio, and other visual cues
- Representation - diverse stock avatars reflecting varied identities
- Dubbing controls - controls to help ensure fidelity in dubbed content
- Video Sharing Permissions - controlled access to keep internal content secure
Community & Enablement
- Synthesia Community - level up your video skills and build lasting connections with other passionate AI video creators.
- Synthesia Academy - educational resources to master video creation with Synthesia
- Futuresafe Report - annual report on responsible and ethical content creation
- Content Integrity Policy - robust framework for reporting and removing harmful or infringing content
- Talent Experience Program - initiative for collaboration and fair compensation for the actors behind stock avatars
- External Red Teaming - independent testing to ensure resilient and secure AI systems
- Dedicated Support Team - live chat and email access for timely customer assistance
- Synthesia Live Events - in-person opportunities to connect, learn, and engage with the Synthesia community
- Adverse Impact Reporting Form - open channel to report concerns related to our AI systems
- Partnership on AI (PAI) - cross-sector collaboration to set standards for responsible AI
- Content Authenticity Initiative - industry engagement to advance verifiable and trustworthy content
Certified to meet global security and compliance standards.
“Large organizations like SAP have high standards for maturity when adopting third-party tools. Synthesia met them effortlessly.”
Head of SAP Experience Garage
AI Governance FAQs.
Is customer data used to train the AI Model(s)?
Synthesia does not use customer data, including inputs or outputs, to pre-train its AI components and any processing of customer data that requires, or results in, the fine-tuning of AI components is based on the written instructions of customers, in accordance with their governing agreement.
Who owns the videos that are created by the service?
The customer retains ownership of the videos. Details will be governed by the Customer Terms of Service.
What controls are in place to prevent the use of Synthesia to create malicious or inappropriate content?
Synthesia has an elaborate moderation procedure in place, to prevent misuse of the Synthesia platform for purposes that violate our Terms and Conditions.
Once the subscription ends, what will happen to the model and data it stores?
The user can at any time request that a video/avatar is deleted, it can then take up to 90 days before it is permanently deleted from our system. If there is a request to delete specific data or delete all video/avatar/voice data on termination of an account, we will delete all copies permanently and provide confirmation of deletion. After termination of an account, if there is no request for deletion, the information will automatically be deleted within 90 days.
How do you ensure the explainability of the AI Solution?
Synthesia is committed to making its AI Components understandable and explainable to customers and their users. The inherent transparency of the Services, where AI outputs are directly tied to customer-provided inputs, ensures that discrepancies between expected and actual outputs are immediately apparent. This clear linkage between input (scripts, voice, likeness) and output (avatar performance) provides a straightforward explanation, allowing customers to both trust and effectively interact with its systems, but also to anticipate harmful scenarios and mitigate abuse.
