This deliverable presents the first version of a structured blueprint that identifies best practices for the responsible development, assessment and deployment of AI systems within the ALFIE project. The document focuses on how ethical principles and European policy requirements can be interpreted and applied in concrete AI application contexts, with particular attention to systems developed or supported through Automated Machine Learning (AutoML). The blueprint is grounded in a desk-based analysis of the European regulatory and ethical landscape for AI, including current and emerging policy initiatives such as the AI Act and the General Data Protection Regulation (GDPR), as well as relevant academic and policy literature on trustworthy AI. Rather than treating these frameworks as abstract guidance, the deliverable examines how their key requirements relate to practical design and evaluation choices in AI systems. Furthermore, a central contribution of the document is the ethics risk assessment of three representative AI use cases used within ALFIE, covering driver state monitoring, automated assessment of web accessibility and corporate compliance analysis. Overall, the deliverable contributes to a clearer understanding of how ethical and legal expectations for AI can be translated into actionable guidance grounded in real application scenarios, supporting the broader goal of developing AI systems that are aligned with European values and societal expectations.
Stella et al. (Fri,) studied this question.