Technical

AI-Generated Content

Content substantially created or modified by generative AI.

How the laws define this

Regulatory frameworks across jurisdictions agree that the concept covers text, audio, image, or video material generated or significantly altered by artificial intelligence systems. The primary point of divergence lies in the degree of alteration or human involvement required for content to fall under regulation. Instruments like Colorado HB 24-1147 and Montana SB 25 set a specific materiality threshold, explicitly excluding minimal edits and requiring that generative AI substantially creates or modifies content to materially alter its meaning to a reasonable person. Similarly, the Media Handbook for Reporting on Artificial Intelligence in Kenya isolates outputs produced primarily by algorithms without substantive human editorial crafting. In contrast, COMELEC Resolution No. 11064 adopts a much wider scope, sweeping in any material created, modified, or manipulated by AI or automated systems—including wholly synthetic works and minor algorithmic alterations—regardless of whether the fundamental meaning has changed.

Synthesised from the 9 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. Showing 6 of the 9 instruments that define this term; all are listed under Definitions below.

Jurisdiction	Instrument	Status	What is distinctive
Colorado, United States	Candidate Election Deepfake Disclosures	In Force	Excludes minimal edits, adjustments, or enhancements that do not materially alter the content's meaning or significance.
Montana, United States	Montana SB 25 - Deepfakes Regulation	In Force	Explicitly excludes content that is minimally altered or where the alteration does not materially change the meaning.
Philippines	COMELEC Resolution No. 11064 (Prohibition of AI-Generated Campaign Content)	In Force	Broadly includes content created, modified, or manipulated using AI tools like chatbots, synthetic voice generators, and image manipulation tools, capturing both wholly synthetic and algorithmically altered materials.
European Union	Code of Practice on AI-Generated Content Transparency	In Force	Emphasizes content where AI plays a fundamental creative or modification role, encompassing a wide array from AI-written articles to deepfake videos.
Kenya	Media Handbook for Reporting on Artificial Intelligence in Kenya (Media Council of Kenya draft)	Adopted	Specifies outputs produced primarily by algorithmic systems without substantive human editorial crafting or intervention, expecting clear labeling and human sign-off.
Uzbekistan	Draft Law on Artificial Intelligence	Proposed	Specifies that AI must be the primary generator rather than a human contributor, subjecting it to mandatory labeling and distribution controls.

What it triggers

In Colorado, disclosure is required for AI-generated deepfakes in election communications with penalties for non-compliance.source
In Montana, knowing distribution of such content as deceptive political deepfakes within 60 days of an election is prohibited.source
Under COMELEC Resolution No. 11064, such content is prohibited in campaigns for the 2025 Philippine elections.source
The European Commission's Code of Practice guides AI providers and deployers in labeling, disclosing, and tracing such content.source

Definitions (9)

Image, video, audio, multimedia, or text content that is substantially created or modified by generative artificial intelligence such that the use of generative AI alters the meaning or significance that a reasonable person would take away from the content; excludes minimal edits, adjustments, or enhancements that do not materially alter meaning.

Colorado HB 24-1147 - Disclosure Requirements for Synthetic Media in ElectionsDefinition 1 of 9

Any text, image, audio, video, or multimedia material created, modified, or manipulated using artificial intelligence, machine learning algorithms, or automated systems, including chatbots, automated reply systems, synthetic voice generators, and image manipulation tools. The term captures both wholly synthetic materials and content altered by algorithmic processes.

COMELEC Resolution No. 11064 (Prohibition of AI-Generated Campaign Content)Definition 2 of 9

Content that is produced or significantly altered by an AI system (including generative AI); under the referenced legal framework, such content must be disclosed to users through notices or technical means (e.g., watermarks) to ensure transparency and proper attribution in public-sector deployments.

Seoul AI-based Administration Enhancement PlanDefinition 3 of 9

Content that has been produced wholly or in part, or materially modified, using artificial intelligence tools or systems; the definition covers text, images, audio, or video outputs where AI played a substantive role in creation or alteration and which may require disclosure to audiences.

Media Guide on the Use of Artificial Intelligence in Kenya (Draft) — Media Council of KenyaDefinition 4 of 9

Outputs produced primarily by algorithmic systems without substantive human editorial crafting or intervention; such content is expected to be clearly labelled and subject to human sign-off, verification and correction procedures under the handbook.

Media Handbook for Reporting on Artificial Intelligence in Kenya (Media Council of Kenya draft)Definition 5 of 9

Content or information that is created wholly or substantially by artificial intelligence systems, including text, images, audio, or video outputs where the AI system is the primary generator rather than a human contributor. Such content is subject to mandatory labeling and distribution controls under the draft law.

Draft Law on Artificial IntelligenceDefinition 6 of 9

Works created by AI systems; refers to creative outputs produced autonomously or with substantial involvement of AI technologies, relevant for assessing copyright ownership and protection.

WIPO Conversation on AI and Intellectual Property PolicyDefinition 7 of 9

Image, video, audio, multimedia, or text content that is substantially created or modified by generative artificial intelligence in a manner that materially alters the meaning or significance that a reasonable person understands from the content. The term does not include content that is minimally altered or where the alteration does not materially change the meaning.

Montana SB 25 - Deepfakes RegulationDefinition 8 of 9

AI-generated content refers to any material, such as text, audio, visual, or audiovisual, that has been significantly produced or altered by an artificial intelligence system. This encompasses a wide array of outputs, from AI-written articles to deepfake videos, where AI plays a fundamental creative or modification role.

Code of Practice on AI-Generated Content TransparencyDefinition 9 of 9

Not to be confused with

Generated or substantially altered using artificial intelligence

That term is a two-part test specifically for AI-created or materially AI-altered media, while this term covers broader substantial creation or modification by generative AI.

AI-generated product

AI-generated product refers to content produced wholly or partly by AI, whereas this term focuses on substantial creation or modification, often with materiality thresholds.

AI-generated/synthetic content

AI-generated/synthetic content emphasizes synthetic production across modalities, while this term includes both creation and modification with varying materiality requirements.

Questions and answers

What counts as AI-generated content under Colorado election law?

Content like images, videos, or text substantially created or modified by generative AI if it alters the meaning a reasonable person would take away, excluding minimal edits. This applies to election communications requiring disclosure.

Does minimal editing with AI trigger disclosure requirements?

No, definitions in Colorado and Montana exclude minimal edits or adjustments that do not materially alter the content's meaning or significance.

What obligations apply to AI-generated content in elections?

Colorado requires disclosure in election communications; Montana prohibits knowing distribution of deceptive deepfakes near elections; Philippines bans it in 2025 campaigns under COMELEC.