PRACTICAL EXAMPLES

When to document or disclose the use of AI

Three common situations show how TransparenteIA turns a real workflow into a clear notice and a traceable record.

Scientific illustration

Assess an AI-generated or AI-edited image used to explain a scientific finding. Record the context, human review and reason for the recommended disclosure.

Advertising image or PDF

Check a synthetic commercial image, create a clear reusable notice and add a visible transparency stamp to the final PDF without uploading the file.

AI-assisted article or report

Document research, drafting or editing assistance, the person responsible for review and the evidence retained even when no public label is required.

From answers to notices: four complete cases

A reproduces the tool’s demo. B is public-interest text with formal checks only; C changes just that answer to substantive review. D leaves the image’s apparent authenticity uncertain. These are synthetic cases evaluated on 2026-09-24.

A · Image

Content: Image
AI use: Partial generation or modification
EU relevance: Yes
Context: Science or research
Real or plausible subject: Yes
Authentic appearance: Yes
Artistic or fictional work: No
Human review: Substantive review by a person with relevant judgement

Disclosure required by the assessed rule

The generated or manipulated content resembles a real or plausible subject or event and could appear authentic. The assessed rule requires disclosure of its artificial nature.

Applied rule: EU-AIA-50-4-DEEPFAKE

Standard

This image was generated or modified with artificial intelligence and may synthetically represent real or plausible elements. A person carried out a substantive review before publication.

Evidence and sources

B · Text

Content: Text
AI use: Partial generation or modification
EU relevance: Yes
Context: News or journalistic information
Public interest: Yes
Editorial responsibility: Yes
Human review: Formal review only: spelling, formatting or style

Disclosure required by the assessed rule

The AI-generated or manipulated text informs the public on a matter of public interest and no substantive human review with editorial responsibility is recorded.

Applied rule: EU-AIA-50-4-PUBLIC-INTEREST-TEXT

Standard

This text was generated or modified with artificial intelligence. No substantive human review with editorial responsibility is recorded before publication.

Evidence and sources

C · Text

Content: Text
AI use: Partial generation or modification
EU relevance: Yes
Context: News or journalistic information
Public interest: Yes
Editorial responsibility: Yes
Human review: Substantive review by a person with relevant judgement

No duty identified under the assessed rules

The public-interest text received substantive human review and a person or organisation assumes editorial responsibility. The specific Article 50(4) duty is not identified, although voluntary disclosure remains available.

Applied rule: EU-AIA-50-4-PUBLIC-INTEREST-TEXT-REVIEWED

Standard

Artificial intelligence was used to generate or modify this text. A person carried out a substantive review before publication.

Evidence and sources

D · Image

Content: Image
AI use: Partial generation or modification
EU relevance: Yes
Context: Science or research
Real or plausible subject: Yes
Authentic appearance: Unsure
Artistic or fictional work: No
Human review: Substantive review by a person with relevant judgement

Review required

A material answer is missing. The tool avoids turning that uncertainty into a legal certainty.

Applied rule: TI-MATERIAL-UNCERTAINTY

Standard

Artificial intelligence was used to generate or modify this image. A person carried out a substantive review before publication.

Evidence and sources

Download answers, rules and notices for all four cases · JSON

Evaluate your case