AI Transparency Checker

Review AI disclosure and marking tasks locally.

AI transparency workspace

Answer questions to see separate Article 50 findings update live. Unknown answers remain visible.

Article 50 transparency screening based on your answers. This is not a full AI Act classification, legal assessment or compliance certification. Implemented means self-reported, not verified. Other AI Act and data-protection duties may apply. Special law-enforcement, standard-editing and artistic exceptions need separate review.

Your use case

Operator role

Providers develop or have an AI system developed and place it on the market or put it into service under their name. Deployers use an AI system under their authority, excluding personal non-professional use. Official role definitions.

Direct interaction

Does the AI interact directly with people?

Is the AI interaction obvious in its context?

Consider a reasonably well-informed, observant and circumspect person, not only your team.

Is an accessible AI notice visible by the first interaction?

Synthetic content

Does the AI generate synthetic audio, images, video or text?

Are outputs machine-readably marked and detectable as AI-generated?

A visible notice alone does not establish machine-readable marking.

Deepfakes

Does the AI generate or manipulate deepfake image, audio or video?

Consider content that resembles real people, objects, places, entities or events and could falsely appear authentic.

Is the artificial origin disclosed by the first exposure?

Public-interest text

Is AI-generated or manipulated text published?

Is that text intended to inform the public on matters of public interest?

Does the content undergo human review or editorial control?

Does a person or organisation hold editorial responsibility?

The text exception needs both this responsibility and human review or editorial control.

Is the artificial origin of the text disclosed by the first exposure?

Biometrics and emotion

Is biometric categorisation used?

Is emotion recognition used?

Are exposed people informed about the system by the first exposure?

Sensitive decisions

Does the AI affect employment or worker decisions?

Does the AI affect creditworthiness or credit access?

Does the AI affect education access or assessment?

Does the AI affect access to essential services?

Live findings

0 review recommended · 6 unknown · 0 self-reported · 0 not applicable

Direct-interaction disclosure

Unknown

Your operator role is unknown; this duty is assigned to the provider.

Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.

Official source: Article 50(1), (5)

Machine-readable synthetic marking

Unknown

Your operator role is unknown; this duty is assigned to the provider.

Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.

Official source: Article 50(2)

Biometric and emotion exposure notice

Unknown

Your operator role is unknown; this duty is assigned to the deployer.

Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.

Official source: Article 50(3), (5)

Deepfake disclosure

Unknown

Your operator role is unknown; this duty is assigned to the deployer.

Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.

Official source: Article 50(4), (5)

Public-interest text disclosure

Unknown

Your operator role is unknown; this duty is assigned to the deployer.

Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.

Official source: Article 50(4), (5)

Further review for sensitive decisions

Unknown

Whether the AI affects one or more sensitive decision areas is unknown.

Next step: Confirm how outputs influence people and arrange a separate risk-classification review where appropriate.

Official source: Article 50(6)

Build an evidence checklist

Select the relevant use cases there. This link does not carry your answers.

Copy or download the answers and all six findings.

Export format

Export preview

AI transparency screening
Article 50 transparency screening based on your answers. This is not a full AI Act classification, legal assessment or compliance certification. Implemented means self-reported, not verified. Other AI Act and data-protection duties may apply. Special law-enforcement, standard-editing and artistic exceptions need separate review.

Role: unknown
Does the AI interact directly with people? unknown
Is the AI interaction obvious in its context? unknown
Is an accessible AI notice visible by the first interaction? unknown
Does the AI generate synthetic audio, images, video or text? unknown
Are outputs machine-readably marked and detectable as AI-generated? unknown
Does the AI generate or manipulate deepfake image, audio or video? unknown
Is the artificial origin disclosed by the first exposure? unknown
Is AI-generated or manipulated text published? unknown
Is that text intended to inform the public on matters of public interest? unknown
Does the content undergo human review or editorial control? unknown
Does a person or organisation hold editorial responsibility? unknown
Is the artificial origin of the text disclosed by the first exposure? unknown
Is biometric categorisation used? unknown
Is emotion recognition used? unknown
Are exposed people informed about the system by the first exposure? unknown
Does the AI affect employment or worker decisions? unknown
Does the AI affect creditworthiness or credit access? unknown
Does the AI affect education access or assessment? unknown
Does the AI affect access to essential services? unknown

Direct-interaction disclosure
Unknown — Article 50(1), (5)
Reason: Your operator role is unknown; this duty is assigned to the provider.
Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Machine-readable synthetic marking
Unknown — Article 50(2)
Reason: Your operator role is unknown; this duty is assigned to the provider.
Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Biometric and emotion exposure notice
Unknown — Article 50(3), (5)
Reason: Your operator role is unknown; this duty is assigned to the deployer.
Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Deepfake disclosure
Unknown — Article 50(4), (5)
Reason: Your operator role is unknown; this duty is assigned to the deployer.
Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Public-interest text disclosure
Unknown — Article 50(4), (5)
Reason: Your operator role is unknown; this duty is assigned to the deployer.
Next step: Establish whether you provide, deploy or perform both roles before relying on this finding.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Further review for sensitive decisions
Unknown — Article 50(6)
Reason: Whether the AI affects one or more sensitive decision areas is unknown.
Next step: Confirm how outputs influence people and arrange a separate risk-classification review where appropriate.
Source: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-50

Answers stay in this tab’s memory. No tool data is uploaded or saved.

Questions

Does a green result prove compliance?

No. Findings reflect your answers within a narrow transparency scope. They do not inspect your application or certify compliance.

Are all publicly published AI texts treated the same?

No. The decision tree distinguishes public-interest publication and human review or editorial control with editorial responsibility. Confirm the actual context against Article 50.

What if I do not know my role?

Choose Unknown. Role-dependent findings stay uncertain and explain which provider or deployer responsibilities to investigate.

Where can I check the applicable rules?

Use the official sources linked in Learn and Reference. Applicable requirements depend on the product, AI system, operator role and actual use.