Unacceptable risk
Banned
These uses are simply not allowed in the EU, whoever builds them.
Social scoring, manipulative systems, untargeted face scraping, most live remote facial recognition
Independent, voluntary AI involvement labels
Who really wrote this? — EU AI Act Content Labels & Transparency Scale
One clear label, five levels — from fully human to fully synthetic. A voluntary way to show how content was made, with plain-language EU AI Act context when it matters.
- 27
- member states covered
- 5
- involvement levels
- In force
- Art. 50 transparency since 2 Aug 2026
AI CONTENT
Involvement label
- Human-authored
- AI-assisted
- Co-created
- AI-generated
- Fully synthetic
AI involvement100% human wording
Human-ledFully synthetic
Human-authored — A person wrote it. AI was only used for spell-check, a thesaurus, or bouncing around ideas. Nothing AI wrote ended up in the final piece.
The scale
Five clear levels of AI involvement
Pick the closest match for how the content was made. These voluntary grades describe provenance — they are not the AI Act's legal risk classes.
C
Co-created
over 50% AI wording, fully reviewed
AI drafts big chunks, a person directs, edits, and fact-checks. A named human or company carries editorial responsibility.
Typical examples
- AI-drafted articles with an editor
- Code generated then reviewed
What the law expects
Human-reviewed public-interest text can be exempt from Article 50(4) when a person or organisation holds editorial responsibility.
Find your label
Three questions. One honest flag.
Get a simple suggested grade for how the content was made — not a legal classification. Nothing is sent anywhere.
1How much of the finished piece did AI actually produce?
2Did a person edit, fact-check, and take responsibility for it?
3Does it show real people, places, or events as if authentic?
Your label appears here
Answer the three questions on the left.
Embed
Put the flag on your work
Use the voluntary badge anywhere: hot-link it on the web, or download it for slides, documents, and PDFs. Every badge links back to a plain-language explanation.
1 · Pick your grade
2 · Pick a format
3 · Pick a size
4 · Copy the snippet
<a href="https://aiflags.eu/label/B" data-ai-label="B">
<img src="https://aiflags.eu/badge/B.svg" alt="AI involvement: B — AI-assisted" height="28" loading="lazy" />
</a>Risk tiers
The AI Act sorts uses, not technologies
The same model can be harmless in a video game and illegal in a police van. Europe regulates the use case, on four levels.
High risk
Strictly regulated
Allowed, but only with testing, documentation, human oversight, and registration.
Hiring and CV screening, exam grading, credit scoring, medical devices, critical infrastructure
Limited risk
Transparency required
Specific uses carry transparency duties so people know when AI is involved.
Chatbots, deepfakes, public-interest AI text, emotion recognition, biometric categorisation
Minimal risk
No risk-tier rules
The Act does not add use-case rules for most low-risk applications. Voluntary codes are encouraged.
Spam filters, recommendation feeds, video game AI, inventory forecasting
The two scales answer different questions.A–E says how a piece of content was made. The legal risk tiers classify an AI system's use case. A content grade never determines a legal risk tier by itself.
Article 50
Six transparency duties, in plain language
Article 50 applies from 2 August 2026. These duties attach to providers or deployers in specific situations — not to every person who uses AI.
- Art. 50(1)
Tell people they are talking to a machine
Chatbots and voice agents must make clear you are interacting with AI, unless it is obvious to any reasonable person.
- Art. 50(2)
Mark synthetic content machine-readably
Providers of systems that generate synthetic text, images, audio or video must make outputs machine-readable and detectable, where technically feasible.
- Art. 50(3)
Disclose emotion and biometric categorisation
If a system reads your emotions or sorts you into biometric categories, you must be informed before it happens.
- Art. 50(4)
Label deepfakes clearly
Deployers must disclose AI-made or manipulated image, audio or video that constitutes a deepfake.
- Art. 50(4)
Flag AI text on public-interest topics
Deployers must disclose AI-generated or manipulated public-interest text unless it had human review or editorial control and someone holds editorial responsibility.
- Art. 50(5)
Make disclosure impossible to miss
Required notices must be clear, distinguishable, accessible, and given by the first interaction or exposure.
Timeline
When each rule starts to bite
The AI Act arrives in waves. Article 50 covers specific transparency cases from August 2026; high-risk system rules follow later.
1 Aug 2024
The AI Act enters into force
Regulation (EU) 2024/1689 becomes law across all 27 member states.
2 Feb 2025
Bans and AI literacy apply
Prohibited practices become illegal. Providers and deployers must ensure staff have sufficient AI literacy.
2 Aug 2025
General-purpose AI rules
Obligations for general-purpose AI models, governance bodies, and penalties start to apply.
2 Aug 2026
Article 50 transparency applies
Defined transparency duties now apply, including chatbot notices and deployer disclosures for deepfakes and certain public-interest AI text.
2 Dec 2026
Grace period for existing generators ends
Providers of generative AI systems already on the market before 2 August 2026 must meet Article 50(2) machine-marking duties.
Upcoming2 Dec 2027
Annex III high-risk use rules
Requirements apply to high-risk uses in areas such as biometrics, education, employment, essential services, law enforcement and migration.
Upcoming2 Aug 2028
High-risk product rules
Requirements apply to high-risk AI used in products covered by the safety laws listed in Annex I.
Upcoming
Print & share
Put it on the wall
Two reference sheets you can hand to a newsroom, a classroom, or a marketing team.
The label
Single card, energy-label stylePNG
The full system
Levels, risks and examples on one sheetPNG
Reference sheets are illustrative artwork for this project. They are not official European Union publications.
Questions
Straight answers
No. aiflags.eu is an independent, plain-language project. The A–E scale is our proposal for making the AI Act's transparency rules easy to apply, modelled on the EU energy label everyone already understands.