AIOLIA Deliverable D3.4 is now ready! It is a mechanism for evaluating Ethics Readiness Levels and algorithmic impact assessment.
AI systems do not fail just because their their benchmarks are low. A system can be technically mature, but ethically immature. Then it performs really well in producing harms that were not anticipated.
Technical readiness is not an all-encompassing goal. It makes sense to talk about ethics readiness in parallel. This is important because design choices become harder to change over time. Ethics must be considered during design, not after deployment. But how can a project manager measure ethics readiness?
AIOLIA mechanism enables a structured interaction between technical AI experts, ethics experts, and domain specialists. In the Ethics Readiness Levels (ERL) tool, ethics-by-design begins with, and proceed through, their dialogue that follows a list of indicators. Ethical reflection happens during this dialogue. AIOLIA ERL mechanism provides a way to make it visible, manageable, open to evaluation, and repeatable across the AI design process.
The ERLs include AIOLIA technical measures and operational guidance from AIOLIA D3.1. Our ethics management tool is informed by real AI use cases, while also being intellectually provocative and eliciting reflection at key points in the AI system design.
The ERL mechanism was presented by Laurynas Adomaitis (RISE) at IEEE ZINC conference on June 3, 2026.