BreAkIng Bad – Hacking the AI Act

• Type of event: Modularized project seminar
• Facilities: Faculty of Mathematics, Informatics & Natural Sciences
• Funding period: 01.09.2024 to 31.03.2025
• Short title: BreAkIng Bad

Orientation of the project

The goal of the seminar “BreAkIng Bad - Hacking the AI Act” was to raise awareness of the risks of (generative) AI and to teach a critical approach towards AI in terms of opportunities and risks before the background of the current legal developments (EU AI Act) in a practical way. As part of this follow-up project, the content developed during the project seminar will be modularized and integrated into existing courses and seminars.

^{deepmind x / unsplash}

Project realisation

The modularization and integration of the project will be achieved as follows:

• Reviewing courses in which the content from BreAkIng Bad can be integrated. Coordination with lecturers regarding the objectives, the students' prior knowledge and the requirements for the teaching materials – the overarching goal is (and remains) to raise awareness for the tension between business opportunities and (ethical and legal) risks when using (generative) AI
• Modularization of the existing concept and materials and further development according to the new format (according to the requirements of the lecturers and based on feedback from the previous seminar)
• Support lecturers in integrating the modular content into their courses

Persons involved

Lucas Memmert

Research Assistant & Project Manager

Focal points:

• Human-AI Collaboration
• Generative/Large Language Models
• Complex Problem Solving

Contact

Dr. Michaela Regneri

Senior Researcher & Project Manager

Research Group: Machine Learning

• Optimization
• Understanding Deep Learning Architectures
• Quantum Computing

Contact

Sandra Timmermann

Student Assistant (B.A. Philosophy, minor in Computer Science)

Focal points:

• Social injustice and the role of IT
• Value-sensitive design
• Neurodiversity, esp. autism