Netflix has released a comprehensive set of guidelines for using Generated AI (Genai) in Global Content Productions, and outlines when and how such tools can be responsibly deployed. The framework is aimed at filmmakers, production partners and vendors, and emphasizes transparency, data security and respect for human resource rights.
This guidance provides production teams with a structured approach to assessing proposed AI applications. While many low-risk cases may not require legal review, any use that involves final deliverables, talent similarities, personal data, or third party intellectual property must escalate for written approval.
Core Guide Principles
Netflix asks partners to evaluate AI use cases against five key principles.
- AI output must not reproduce or substantially replicate the similarity of copyrighted materials or ownership.
- Tools must not store, reuse or train the production inputs or outputs of the platform.
- Whenever possible, the generation tool should work in an enterprise-configured environment.
- AI-generating materials are considered temporary materials and should be excluded from the final deliverable unless cleared.
- Genai should not be used to exchange or generate human resource performance or union-covered work without consent.
If your use case meets all these criteria, it is enough to provide ample notification of platform contacts. However, uncertainties and exceptions require escalation and approval.
Use cases that require written approval
Guidelines always specify scenarios where written consent is required.
- Data usage: AI tools must not provide unreleased Netflix assets, scripts, images or personal information such as cast or crew without explicit clearance. A training model for third-party materials also requires legal rights.
- Creative output: Do not use AI to generate key story elements such as main characters and central visual designs without approval. Referencing copyrighted works, public figures, or portraits of real estate control is restricted.
- Human resources and performance: Significant changes to performance, such as changes in digital replicas or delivery or changes in emotional tone, require documented consent and compliance with union rules.
- Ethics and Expression: Content generated by AI should not mislead audiences, produce events, or replace work on behalf of unions to avoid agreement.
Confidentiality and Data Protection
To protect sensitive production data, the platform recommends using AI tools covered by enterprise agreements. This prevents input training and resale. Still, talent similarities, unpublished footage, or use of confidential material must escalate for approval.
Final output and temporary media
This guidance distinguishes between temporary AI-generated mockups and materials that appear in the final artifact. Exploratory use during creative development is permitted, but visual, audio, or text elements generated in AI reaching the screen, whether in the foreground or background, can cause clearance requirements.
Human resource strengthening considerations
Special regulations apply when using AI to digitally enhance or replicate performance. You must agree to digital replicas. However, if the performer is not recognised, there are only exceptions in limited contexts such as re-shooting, safety-related depictions, and scenarios. Standard post-production changes (e.g., continuity corrections, cosmetic adjustments, sound clarity) are still acceptable. It also directs that AI-trained models for human resource manipulation are production-specific and should not be reused throughout the project without consent.
Vendor and Workflow Monitoring
When production partners work with vendors employing custom AI workflows, each component must meet the platform's requirements for data protection, consent, and creative integrity. The same standards apply whether the work is carried out internally or outsourced.
