Students will learn:
- why different families of AI (symbolic, machine learning) present different risk profiles
- how risks of bias and discrimination manifest themselves in AI systems, and approaches to mitigate those risks
- why explainability prevents machine learning from being used in certain critical applications
- "meaningful human control" of AI systems
- the legal framework around AI: the European General Data Protection Regulation, Platform Regulation, the future AI Act, EU Charter of Fundamental Rights
Teaching will be structured around use cases, including autonomous lethal weapons, facial recognition, social media use of algorithms, algorithms to detect terrorist risk
Grading will be half based on multiple choice test, and half on a final paper.