The architecture of the NEANIAS project proposes a view in which thematic services are developed on top of more general purpose Core Services, which enable the usage of resources of the computational environment in which they are deployed, within the context of the European Open Science Cloud.

Artificial Intelligence, in general, refers to a whole discipline, encompassing a variety of different approaches, that can hardly be mapped to a service or set of services, however complex and powerful. Nonetheless, the successes of Machine Learning techniques and the growth of interest in their application in the most diverse fields, a process sometimes referred to as the Deep Learning revolution, has made it clear that some AI approaches and techniques could be fruitfully engineered as services of general interest within a cloud environment. There are whole projects aimed at achieving this kind of result, even outside the EOSC context (see, e.g., the Horizon 2020 funded AI4EU initiative), but within NEANIAS we wanted to explore the chance to pursue this objective with the concrete perspective of supporting thematic, open science services, at the same time empowering researchers within the Consortium, as well as users of these services, but also presenting successful case studies of AI applications to open science (and guidance on how to develop additional ones) for further developments within EOSC. AI services developed within NEANIAS, therefore, will be employed firstly within NEANIAS, but they will be made available to the EOSC just like thematic services.