Administrative Information
| Title | AI Risks and Risk Management |
| Duration | 90 |
| Module | B |
| Lesson Type | Interactive Session |
| Focus | Ethical - Trustworthy AI |
| Topic | Risk |
Keywords
AI Risk,Risk Management,Resilience Management,Explainable AI,Interpretable AI,Security,Resilience,Transparency,Accountability,
Learning Goals
- Recognize, analyze and discuss the necessity and the role of security testing in AI.
- Employ documents, such as the AI Act, to classify use cases into risk categories.
- Discuss in-depth and with common European vocabulary on AI risk and how can AI be designed safely and benevolently, in ways that will not expose humans to unacceptable risks of harm, even when risky behaviour may advance a system goal.
- Classify the obligations of providers and users in relation to AI risk management.
- Discuss the requirements for high-risk AI systems as per the AI Act.
- Elaborate and recognize specific risks and risk taxonomies, relevant throughout the different stages of the AI lifecycle.
Expected Preparation
Learning Events to be Completed Before
None.
Obligatory for Students
- High-Level Expert Group on Artificial Intelligence (AI HLEG). (Apr. 8, 2019). European Commission. “Ethics Guidelines for Trustworthy AI.
- HCAIM Lecture on Risk and Risk Mitigation
Optional for Students
- HCAIM Webinar on the European Approach Towards Reliable, Safe, and Trustworthy AI (Available on YouTube)
- HCAIM Webinar on the Role of Security and Privacy in Machine Learning (Available on YouTube)
References and background for students
- High-Level Expert Group on Artificial Intelligence (AI HLEG). (Apr. 8, 2019). European Commission. “Ethics Guidelines for Trustworthy AI.
- HCAIM Lecture on Risk and Risk Mitigation
Recommended for Teachers
- European Commission: Proposal for a Regulation of the European Parliament and of the Council Laying Down Harmonized Rules on Artificial Intelligence (Artificial Intelligence Act), 2021
- European Union Agency for Cybersecurity, Malatras, A., Dede, G.,AI cybersecurity challenges: threat landscape for artificial intelligence, European Network and Information Security Agency, 2020
- High-Level Expert Group on Artificial Intelligence (AI HLEG). (Apr. 8, 2019). European Commission. “Ethics Guidelines for Trustworthy AI.”
- Malicious Uses and Abuses of Artificial Intelligence 2020, Europol Publication
Lesson materials
Instructions for Teachers
This interactive session aims to build a common vocabulary for students to discuss risk and risk management related to AI, specifically through the EU risk-based approach towards AI. The materials provided, as well as the overall outline and schedule of the interactive session are indicative and presented in order to better guide a discussion with the students. As long as the learning goals are met, instructors are encouraged to follow the natural flow of the discussion.
Outline and Schedule of the Interactive Session:
| Duration | Topic | Description |
|---|---|---|
| 15 mins | Introduction | Defining risk. EU's approach towards AI risk management. Risk management strategies throughout the AI lifecycle. |
| 30 mins | Classifying risk | AI assets and threat taxonomy and risk management frameworks to describe technical design characteristics for AI. |
| 30 mins | Mitigating risk | Approaching risk mitigation using use cases and common vocabulary, and reference models to leverage the discussion. |
| 15mins | Closing Remarks | Discussion with Students. Questions and Answers. |
Two HCAIM Webinars are particularly relevant to this interactive session. Those could be shown prior to the discussion or parts of them could be played to facilitate topics and manage silence.
Acknowledgements
This presentation was developed by Dr George Sharkov and Christina Todorova at the European Software Institute – Center Eastern Europe.
The presentation uses slides from a presentation by Dr Mihály Héder at the Budapest University of Technology and Economics. Please, consider following his work.
The Human-Centered AI Masters programme was Co-Financed by the Connecting Europe Facility of the European Union Under Grant №CEF-TC-2020-1 Digital Skills 2020-EU-IA-0068.