Repeated iteration
A project reaches legal, security or procurement without the evidence those functions need, then returns to the team for another build cycle.
Loading
Please wait...
EU AI Act guide, updated 30 August 2026
What the EU AI Act requires now, and what can wait
The AI Act is not one deadline or one checklist. Your next action depends on whether you provide or deploy the system, what it does, who it affects and which part of the amended timetable applies.
AI literacy
Applied since 2 February 2025
Article 4 requires providers and deployers to take measures supporting staff AI literacy in context.
General application
Most provisions from 2 August 2026
GPAI, governance and enforcement provisions have their own dates and transition rules.
High-risk requirements
Annex III: 2 December 2027
Annex I product-safety high-risk systems follow on 2 August 2028.
AI literacy and Article 4
Train for the decisions people actually make
Article 4 requires measures that support AI literacy while taking account of technical knowledge, experience, education, training and the context in which AI is used. It does not prescribe one sufficient level, a specific course or a mandatory certificate.
A generic awareness session cannot resolve what each team may share, what it must verify, when it should escalate or where the tool's autonomy ends.
- Map the AI systems and use cases people actually handle
- Define role-specific decisions, safeguards and escalation paths
- Use practical scenarios to test understanding, not attendance alone
- Keep internal records of training and other guidance measures
- Review the programme when tools, roles or risk context change
A Governance School completion certificate records who completed the course and assessment, and when. It can support a wider Article 4 programme, but it does not by itself establish compliance.
Read the current European Commission AI literacy Q&AOperational readiness
Five decisions that prevent late rework
The work starts with the system and operator role, not with a generic template.
- 01
What role does the company hold?
Separate provider, deployer, importer and distributor responsibilities. A company can hold more than one role across its AI portfolio.
- 02
Is the use prohibited, high-risk or subject to transparency rules?
Classify the actual intended purpose, affected people and deployment context. Do not infer the category from the model name alone.
- 03
What evidence will let this use move forward?
Agree the validation, data, documentation and human oversight needed before launch so legal and technical review do not restart at every iteration.
- 04
Who decides, escalates and monitors?
Name accountable owners and intervention paths that work in the real workflow, not only in a policy document.
- 05
Which date and transition rule applies?
Use the amended Article 113 timetable and any legacy-system provision that applies to the specific system.
Current application timetable
The dates companies should separate
2 February 2025
Article 4 and most prohibited-practice rules
AI literacy measures already apply. Specific newly added prohibited-practice provisions apply from 2 December 2026.
2 August 2025
GPAI, governance and penalty framework provisions
Chapter V and specified governance and enforcement provisions began to apply, subject to transition rules for older GPAI models.
2 August 2026
Most remaining provisions
The Regulation generally applies, with material exceptions for high-risk requirements and selected amended provisions.
2 December 2027
Annex III high-risk requirements
Chapter III Sections 1 to 3 apply to high-risk systems classified under Article 6(2) and Annex III.
2 August 2028
Annex I product-safety high-risk requirements
The same sections apply to high-risk systems classified under Article 6(1) and Annex I.
Business consequence
The cost appears before the regulator does
Lost confidence
Customers and leadership lose trust when the company cannot explain who approved an AI use, what was tested or how staff were prepared.
Investment without adoption
Employees avoid useful AI or use it inconsistently because policy language never became a practical workplace decision.
Choose the problem
Start where the uncertainty is costing you
Each route turns one part of the AI Act into an operational decision.
Course ยท 01
AI Literacy in Practice
Give every team one practical baseline, assessment and completion record.
Service ยท 02
High-risk classification
Resolve role, intended purpose, Annex mapping and the next evidence required.
Service ยท 03
AI use policy
Turn data, tool, verification and escalation rules into decisions staff can follow.
Policy-to-Practice AI Enablement Pilot
Prove the approach on one real use case
Choose one process, team or critical AI situation. We turn its policy, system and risk context into practical training, implement the pilot and return participation and completion evidence before you consider a wider rollout.
- 01
Choose the decision
Identify one use case or recurring review question where inconsistent interpretation is creating delay, exposure or lost value.
- 02
Build the micromodule
Translate the relevant policy, system and risk context into a short scenario with a decision, feedback and escalation path.
- 03
Run and evaluate
Deploy to the pilot group, review participation and completion, and decide what should scale, change or stop.
Bring one AI use case. See how it becomes practical training.
Share one AI use case or policy challenge. We will show how Governance School would turn it into a workplace micromodule before you consider a wider rollout.
Founder, AIGP-certified AI governance professional, compliance lawyer and software engineer
Jhonathan Campos
Jhonathan Campos develops Governance School's AI governance and compliance content at the intersection of law, software engineering and operational implementation. His work focuses on translating regulatory requirements, technical constraints and company policy into decisions that teams can apply.
Primary sources
Current law, not a recycled timeline
Booking
Get in touch
Share your goals with us and discover how we can guide you through complex compliance requirements.