Technical

Synthetic content

Media created or substantially altered by AI.

How the laws define this

Regulatory frameworks agree that synthetic content consists of media created or materially altered through artificial intelligence or algorithmic processes rather than directly captured from real-world events or created purely by humans. They unite in targeting substantive automated generation that affects content authenticity or provenance. However, the instruments diverge on covered media types and explicit exclusions for minor edits. The Deepfakes Guidelines (SDAIA) and an act to amend the general business law, in relation to requiring synthetic content creations system providers to include provenance data on synthetic content produced or modified by a synthetic content creations system that the synthetic content creations system provider makes available explicitly limit their scope to audio or visual media while carving out routine or minor modifications like stylistic filters, restoration, and red-eye tools. Conversely, the Content Origin Protection and Integrity from Edited and Deepfaked Media Act (S. 1396, 2025) applies broadly to content generated wholly or primarily by algorithmic processes, omitting explicit minor edit exclusions.

Synthesised from the 3 statutory definitions below. Each one is quoted in full, with its source.

How jurisdictions differ

What is distinctive in each definition, binding law in force first.

Jurisdiction	Instrument	Status	What is distinctive
New York, United States	An act to amend the general business law, in relation to requiring synthetic content creations system providers to include provenance data on synthetic content produced or modified by a synthetic content creations system that the synthetic content creations system provider makes available	Adopted	Explicitly excludes minor non-material changes like red-eye filters, narrowing to audio or visual content materially modified or generated by a synthetic content creations system.
United States	Content Origin Protection and Integrity from Edited and Deepfaked Media Act (S. 1396, 2025)	Proposed	Widens to content generated wholly or primarily by algorithmic processes including AI, not limited to audio or visual, and ties to potential provenance and disclosure requirements.
Saudi Arabia	Deepfakes Guidelines (SDAIA)	Draft	Distinguishes substantive AI-generated or transformed media from routine benign edits like restoration or stylistic filters, emphasizing departure from original captured reality.

What it triggers

In New York, United States, synthetic content creations system providers must include provenance data on synthetic content they produce or modify and make available.source
Under the United States Content Origin Protection and Integrity from Edited and Deepfaked Media Act (S. 1396, 2025), synthetic media tool providers would be required to embed content provenance metadata.source
The Saudi Arabia Deepfakes Guidelines (SDAIA) set out practical guidance on the responsible creation and labelling of synthetic content for developers, platforms, content creators and public-sector entities.source

Definitions (3)

Media (audio, image, or video) that is created or materially altered using artificial intelligence techniques such that its content or appearance departs from the original captured reality. The term distinguishes routine/benign edits (e.g., restoration, stylistic filters) from substantive AI-generated or AI-transformed media that may affect attribution or authenticity.

Deepfakes Guidelines (SDAIA)Definition 1 of 3

Content that is generated wholly or primarily by algorithmic processes (including AI) rather than being directly captured from real-world events or human creators, and that may be subject to provenance and disclosure requirements.

Content Origin Protection and Integrity from Edited and Deepfaked Media Act (S. 1396, 2025)Definition 2 of 3

Audio or visual content that has been generated or materially modified by a synthetic content creations system. This definition is crucial for distinguishing between genuinely human-created content and that which has been significantly altered or produced by AI, explicitly excluding technologies like red-eye filters or other similar tools that are only capable of making minor, non-material changes to existing audio or visual content.

An act to amend the general business law, in relation to requiring synthetic content creations system providers to include provenance data on synthetic content produced or modified by a synthetic content creations system that the synthetic content creations system provider makes availableDefinition 3 of 3

Not to be confused with

Synthetic media

Synthetic media focuses specifically on AI-generated or materially altered image, audio, or video, while synthetic content definitions vary in scope across the three.

AI-generated/synthetic content

AI-generated/synthetic content covers information produced by AI across all media modalities, whereas these definitions center on audio, visual, or algorithmic media.

Questions and answers

What minor edits are excluded from synthetic content under New York law?

Technologies like red-eye filters or similar tools that only make minor, non-material changes to existing audio or visual content are explicitly excluded.

Does synthetic content under the US bill include non-AI algorithmic processes?

Yes, it covers content generated wholly or primarily by algorithmic processes including AI, rather than directly captured from real-world events or human creators.

What does Saudi Arabia's guidelines distinguish synthetic content from?

Routine or benign edits such as restoration and stylistic filters, focusing instead on media where content or appearance departs from original captured reality due to AI.