EU AI Act Q1 2025: Phased Deadlines & Annex III Education Classification
The EU AI Act is rolling out in phases through 2027. If your AI touches education, credentialing, or student profiling—here's what you need to prepare now.
Pulse Insight
The EU AI Act is being rolled out in phases with material deadlines that materially affect deployment risk. If your model will be used in classroom assessments, adaptive learning, credentialing, or student profiling—assume high-risk classification and start preparing compliance artifacts now.
Key Phased Deadlines
The AI Act entered into force 1 Aug 2024, with phased applicability through 2 Aug 2027:
2 Feb 2025
Initial provisions take effect—prohibitions on unacceptable-risk systems plus basic obligations.
2 Aug 2025
General-Purpose AI (GPAI) obligations apply, including transparency and documentation duties for providers.
2 Aug 2026
High-risk AI rules become enforceable, including Annex III categories (education among others). High-risk systems must meet robust governance, risk management, human oversight, and conformity assessment requirements.
2 Aug 2027
Full compliance horizon for AI integrated into regulated products, and providers of GPAI models placed on market before 2025 finalize compliance.
Annex III: Education as High-Risk
Annex III explicitly lists use cases qualifying as high-risk, including AI systems used in education. This triggers extensive obligations once the 2026 timeline hits.
High-risk educational AI includes:
- Classroom assessments and grading
- Adaptive learning systems
- Credentialing and certification
- Student profiling and tracking
- Admissions and selection systems
What to Act on Now
If your model intersects with education, prepare these artifacts by early 2026:
- [ ] Technical documentation covering system design, training data, and intended purpose
- [ ] Risk assessment identifying potential harms and mitigation measures
- [ ] Human oversight design ensuring meaningful human control over high-stakes decisions
- [ ] Conformity assessment artifacts demonstrating compliance with essential requirements
- [ ] Transparency documentation for users and affected persons
- [ ] Data governance records covering training and testing data provenance
Practical Compliance Signals
For GPAI providers (Aug 2025):
- Prepare model cards and technical documentation
- Document training data sources and methodologies
- Establish transparency mechanisms for downstream deployers
For high-risk deployers (Aug 2026):
- Implement quality management systems
- Establish post-market monitoring
- Ensure logging and traceability
- Design meaningful human oversight mechanisms
The Governance Imperative
The EU AI Act treats AI governance as a continuous obligation, not a one-time certification.
Key ongoing requirements:
- Regular risk reassessment
- Incident reporting
- Documentation updates
- Human oversight verification
- Conformity maintenance
Disclaimer
Informational only. This article provides general guidance on EU AI Act requirements and should not be construed as legal advice. Consult qualified legal counsel for specific compliance obligations.