The S3 Observatory has been built in collaboration among DG REGIO G1, DG REGIO country-desks and the S3CoP Secretariat. The project team identified S3 priorities directly from regions' and member states' S3 documents and deployed tools from Artificial Intelligence (AI), including Generative AI and entity extraction and disambiguation for automatic topic tagging and classification. AI supported the following exercises:
In particular, AI tools supported the classification of S3 priorities across NACE Sections and divisions, NABS codes (1 and 2 digits), and Industrial Ecosystems. A full description of the methodology is available here.
An essential disclaimer for this analysis is that AI is not error-free, as AI models have limitations in contextual comprehension. To address this limitation, we incorporated human supervision and quality review to maximise the accuracy and consistency of the results. This synergy between generative AI and human expertise enhances the system's reliability. Nevertheless, as a pioneering exercise, the S3 Observatory may result in eventual misclassifications or inaccuracies due to the inherent complexity of natural language understanding. For the reasons above, the S3 Observatory will undergo periodic reviews.
Regional/national institutions or bodies, responsible for the management of the smart specialisation strategy can contact us at contact@s3-cop.eu to rectify or integrate the information provided.
| Region name (NUTS ID) | Priority | Strategy | |
|---|---|---|---|
