EU AI Act — Compliance
EU 2024/1689Last updated: November 2026
TentaGen is designed in accordance with the EU Artificial Intelligence Act (EU AI Act, 2024/1689). Here we describe how we meet the requirements.
What is the EU AI Act?
The EU AI Act is the world's first comprehensive legislation for artificial intelligence. It entered into force in phases from August 2024 and regulates how AI systems may be developed, distributed, and used within the EU. Its purpose is to ensure that AI is used in a safe, transparent, and fair manner that respects fundamental rights.
Classification of TentaGen
The AI Act identifies AI systems in education as potentially high-risk (Annex III, point 3). Under Article 6(3), however, AI systems that perform a preparatory task for an activity assessed by a human are exempt, provided the system does not pose a significant risk. TentaGen meets condition (d) of Article 6(3): the tool generates question suggestions that are always reviewed and approved by an educator before use. No question reaches a student without human judgment. The system makes no autonomous decisions and does not affect students' rights.
Transparency (Article 50)
In accordance with Article 50, we clearly inform that: TentaGen uses a language model from OpenAI to generate exam questions. Anthropic's Claude may check questions that have numerical answers or are labelled at a higher taxonomy level, and flag answer keys that may be wrong and levels that seem overstated. Claude also reads and describes images you upload as material and runs Fact-check, which can test a question against material you uploaded or pasted, sources on the web, or both. All generated questions are AI-generated suggestions that require human review. Educators have full control over which questions are used in examinations. When a teacher exports questions that TentaGen has written, every file states that they are AI-drafted. When a teacher converts their own questions to WISEflow, TentaGen adds no marking. The Export Options section of the documentation shows where the marking appears in each format.
Human Oversight
Although TentaGen is not classified as high-risk under Article 6(3), we have adopted human oversight as a design principle: No question reaches a student without an educator having reviewed and approved it. Educators can edit, refine, or reject all AI-generated suggestions. The system requires active action from the educator — questions are never published automatically. Review history is logged for institutional follow-up.
Data and Privacy
We comply with GDPR for all data processing. Uploaded files are deleted from our storage once their text has been read, and images that belong to questions are stored on our servers. The AI providers keep what they receive for a time under their terms, and under OpenAI's and Anthropic's API terms it is not used to train their models. Full details are available in our Privacy Policy.
Timeline and Registration
The EU AI Act applies in phases: February 2025 — prohibited AI practices and AI literacy requirements (Article 4); August 2025 — obligations for general-purpose AI models; August 2026 — the transparency rules (Article 50); December 2026 — the marking of AI-generated content (Article 50(2)) in systems on the market before 2 August 2026 (Article 111(4)); December 2027 — requirements for high-risk AI in education (Annex III), after the EU's Digital Omnibus on AI postponed them. TentaGen is assessed as not high-risk under Article 6(3), and the system will be registered in the EU AI database in the coming months.
Contact
Questions about TentaGen's AI compliance? Contact us at support@tentagen.se.