The field of scientific AI/ML has evolved rapidly in recent years. As AI/ML implementations are applied to complex life science challenges, common issues have emerged, such as the output of poorly described AI datasets and models that commonly do not follow best practices. This has created fundamental problems with reproducibility, explainability, and trustworthiness, limiting the value of published AI datasets and models for the wider scientific community. In response, a diverse range of new AI/ML standards, metadata schemas, and AI-focused resources such as registries have been developed. However, these efforts have led to fragmentation and poor interoperability between AI/ML support infrastructure both within and outside of ELIXIR, creating a new set of challenges for building a harmonised ecosystem of impactful and reusable AI/ML research outputs.

The AI Ecosystem Focus Group brings together experts from ELIXIR’s Nodes  to address growing challenges within the scientific AI/ML field, particularly  the need for harmonisation in AI/ML infrastructure. Building on the foundational work of the ELIXIR Machine Learning Focus Group, the AI Ecosystem Focus Group’s main purpose is to coordinate efforts, address fragmentation, and support the standardisation of AI/ML infrastructure, ultimately aiming to build a more impactful and reusable ecosystem for AI/ML research outputs across ELIXIR.