General Purpose AI (GPAI) models are very popular and implemented in AI systems throughout all industries. Especially large language models (LLMs) with their wide range of applications gained attention, due to their versatility and capability. With the EU AI Act, engineering teams need to think beyond just performance and start considering compliance, reliability, and risk mitigation when using those models.
This session provides a clear introduction to the GPAI Code of Practice, a guideline for ensuring compliance with the EU AI Act’s obligations. We will break down its impact on model providers and downstream providers, distinguishing between modifiers and integrators, outlining their specific responsibilities along the AI Lifecycle.
Additionally, we will introduce practical tools and methods to seamlessly integrate compliance into the machine learning development workflow: from data and model evaluation, to deployment and system maintenance evaluation. A special focus will be given to recommended tools and methods for high-risk AI systems.
Who Should Attend?
This training is designed for engineering teams and other tech-savvy actors looking to use GenAI models in their products while ensuring compliance with the EU AI act.
What prior knowledge is required for the event?
No prior legal knowledge of the EU AI Act is required. Some engineering background is beneficial.
What Are the Expected Learning Outcomes?
After this session, you will be able to:
- Recognise the purpose, key principles, and relevance of the GPAI Code of Practice in AI compliance.
- Identify common safety and security challenges in GPAI.
- Describe the key compliance obligations for GPAI model providers and for downstream providers who are modifying or integrating these models into AI systems.
- Recognise key methods and tools for evaluating data, models, deployment, and system maintenance in GenAI development.
- Identify key compliance requirements for GPAI systems in high-risk use cases and name methods and tools for risk mitigation.
Join us to gain actionable insights on building trustworthy GenAI systems that comply with the latest EU AI regulations!
Trainer*innen: AI Regulatory Experts, appliedAI Institute
* Dieses Training wurde vom Bayerischen Staatsministerium für Digitales gefördert.
