AI Transparency Notice

DRAFT. This text has been written but has not yet been reviewed by a lawyer. The same banner is shown inside the app. It will be removed once legal review is complete.

This notice is the mandatory disclosure required by art. 50 of the European Union Artificial Intelligence Act (AI Act).

1. YOU ARE INTERACTING WITH AN AI SYSTEM

The "agents", "specialists" and "manager" in this app are not real people. All of them are text produced by large language models. Sounding human does not make them human.

2. GENERATED CONTENT IS AI OUTPUT

Reports, analyses, tables, images and suggestions are produced by AI. Stating this in content you share externally is your responsibility.

3. IT CAN BE WRONG — KNOWN LIMITS

It can produce untrue information and present it convincingly
It can make mistakes in numeric calculations
It does not know events after its training cut-off
It can misread a source it has been given
Text coming from a system you connect can look like an instruction to the model and try to steer it (prompt injection)

4. HUMAN OVERSIGHT IS REQUIRED

Every action with lasting effect is put to you for approval. No automated decision with legal effect is taken without your approval (GDPR art. 22).

5. HIGH-RISK AREAS

This app is NOT DESIGNED TO BE USED AS A DECISION TOOL in recruitment, credit assessment, insurance, educational assessment, medical diagnosis, judicial proceedings or law enforcement. If you use it in these areas, the high-risk system obligations under the AI Act pass to you.

6. EVIDENCE CHAIN

Every source your agents read is recorded, and the model's citations are compared against the actual reads; citations that cannot be verified are flagged. This mechanism reduces error — it does not remove it.

7. FEEDBACK

If you come across wrong or harmful output, you can report it to us.