EU AI Act - Running high-risk AI without breaking the rules - Intermediate

Technology & AI - Regulatory compliance
This course is for the teams who build, buy, and operate high-risk AI systems day to day. It turns the EU AI Act's high-risk obligations into a working routine: mapping a system to the right risk tier, keeping the Annex IV technical documentation, designing human oversight, governing data and bias, logging and reporting incidents, meeting GPAI duties, and getting through conformity assessment and the CE mark.

By the end, you'll be able to map a system to the four risk tiers, build and maintain the Annex IV technical documentation, design and prove Article 14 human oversight, govern data under Article 10, run Article 12 logging and Article 73 incident reporting, apply the GPAI duties, and work through the conformity assessment routes and the CE mark for AI.

Also in the EU AI Act series:
← Foundations: What the AI Act means at your desk
→ Advanced: EU AI Act: Advanced Compliance Training

Access

Self-Paced Learning

Lessons

11 Lessons

Certification

Certificate of Completion

Price

€300
Learning Objectives
Course Benefits
Modules
By the end of this course, learners will be able to:

  • Map an AI system to the four risk tiers of the EU AI Act and explain what each tier triggers
  • Build and maintain the Annex IV technical documentation and the conformity dossier for a high-risk system
  • Run Article 12 logging and Article 73 incident reporting: recognise a reportable incident, apply the threshold, and meet the timeline
  • Design and evidence Article 14 human oversight for a high-risk AI system
  • Govern data and address bias under Article 10 so a system stays compliant through change
  • Apply the GPAI duties and work through conformity assessment and the CE mark for AI
Benefits for the Organisation

• Reduce the most common high-risk AI failings - outdated technical documentation, a missed incident report, weak human oversight, ungoverned data - before they become an enforcement notice
• Turn the Act's high-risk obligations into a repeatable weekly and monthly routine, so nothing is left to memory or chance
• Evidence completed AI Act training to regulators, auditors, insurers, and customers with timestamped records
• Standardise the language and escalation route the whole organisation uses, with no improvised handling of an AI compliance concern
• Save your compliance leads and managers the time spent rescuing missed risks and incomplete AI records

Benefits for Learners

• Know exactly what to do in the first 48 hours when a regulator opens an enquiry or a serious AI incident occurs
• Keep live technical documentation and logs that stand up to a conformity assessment
• Handle an AI compliance concern with the confidence to record what you saw and escalate it to the right person
• Design human oversight that meets the Article 14 duty, not just a tick-box
• Pass the assessment knowing every answer traces to the EU AI Act, its annexes, or official regulator guidance, not a study guide

Lesson 1: Welcome and how this works
Meet the people this course is for, see how the lessons fit together, and set out what you will be able to do by the end. Includes the training disclaimer, the EU, UK, and Malta scope, and the as-of date.

Lesson 2: Mapping a System to the Four Risk Tiers
Learn the four risk tiers - unacceptable, high, limited, and minimal - and how to place a real system in the right one. Work through classifying an example system and what its tier triggers.

Lesson 3: Annex IV Technical Documentation
Build the Annex IV technical documentation: what it must contain, the conformity dossier behind it, and how it is kept live. Work through documenting a high-risk system against the Annex IV requirements.

Lesson 4: Article 14 Human Oversight
Design human oversight under Article 14: the measures that let a person understand, monitor, and step in on a high-risk system, and how to prove they work. Work through specifying oversight for an example deployment.

Lesson 5: Article 10 Data Governance and Bias
Govern data under Article 10: relevance, representativeness, and the bias problem in training, validation, and testing data. Work through spotting and addressing a bias risk in an example dataset.

Lesson 6: Article 12 Logging and Article 73 Incident Reporting
Operate Article 12 automatic logging and Article 73 serious-incident reporting: what must be logged, what triggers a report, the timeline, and the reporting route. Tell a routine log apart from a reportable incident with worked cases.

Lesson 7: GPAI Duties and the Code of Practice
Understand the duties on general-purpose AI models, the systemic-risk threshold, and the GPAI Code of Practice. Work through what a provider and a deployer of a GPAI model must each do.

Lesson 8: Conformity Assessment and the CE Mark
Work through the conformity assessment routes for a high-risk system, the declaration of conformity, and the CE mark for AI: what each step means and how to complete it. Work through taking an example system to CE marking.

Lesson 9: Resources
A working reference library: the verified EU AI Act text and annexes, official EU and national regulator guidance, and further reading. All links verified at build.

Lesson 10: Final assessment
Ten multiple-choice questions across the risk tiers, technical documentation, human oversight, data governance, logging and incident reporting, GPAI duties, and conformity. Eighty per cent to pass, two attempts, and a certificate on completion.

Lesson 11: Provide Feedback and Claim Certificate
Share your feedback on the course and download your certificate of completion.

CORPORATE TRAINING

We Can Train Your Employees

Every organisation is different - and so are its learning needs.

Request a call back to talk through:


  • Your current onboarding or training challenges
  • Skills gaps across teams or roles
  • Compliance, wellbeing, or development priorities
  • How digital learning can support performance and retention
  • Special pricing options for multiple seats or team-wide access


We’ll help you explore practical, people-first solutions that fit your organisation.

Course FAQ's

Who is this course for?

The teams who build, buy, and operate high-risk AI systems day to day, and the operations staff who support them. It assumes Foundation-level awareness of the EU AI Act and builds the day-to-day operating skills on top.

This is an intermediate course. It assumes Foundation-level awareness of the EU AI Act; no other prior knowledge is needed.

Where and when can I take this course?

The course can be taken at any time and from any internet-connected device.

Do I need prior knowledge or experience?

This is a foundation-level course, and all concepts are explained in a clear, simple, and practical way. Learners, do not require any previous training or background knowledge.

Can this course be used for company-wide training?

Yes. It is built for the teams who operate high-risk AI systems, and can be rolled out across every team that builds, buys, or runs one.

Do I get a discount if I buy multiple seats for this course?

Yes. Please contact us and we will issue a quote according to your needs.

Will managers be able to track progress and completion?

Yes. Managers and administrators can track learner progress, completion rates, and assessment performance across the course.

This gives clear visibility on engagement and outcomes, making it easier to monitor participation, support staff where needed, and evidence completed AI Act training to regulators and your insurers.

Is this course part of a larger AI compliance programme?

Yes. This is Course 2 (Intermediate) of the Auren EU AI Act bundle - the operating course for the teams who run high-risk AI day to day. It follows Course 1 (Foundation), the awareness floor for everyone who uses or buys AI, and is followed by Course 3 (Advanced), for the senior managers, DPOs, and leaders who carry the legal and personal liability.

All three courses in the bundle can be taken as stand-alone courses.