Draft ToC of the study

Draft ToC of the study on data sovereignty of AI algorithms for Data Spaces

1 Executive summary

1.1 Main data

1.2 Other data

2 Alignment with existing policies

3 Methodology

4 Selected Experts

4.1 Other

5 Dimensions of classification

5.1.1 By location of the algorithm

5.1.1.1 On premises

5.1.1.2 Cloud Based

5.1.1.3 Federated Learning

5.1.2 By access to individual registries

5.1.2.1 Centralized Algorithms

5.1.2.2 Decentralized Algorithms

5.1.3 Transparency and explainability

5.1.3.1 Black box algorithms

5.1.3.2 Glass box algorithms

5.1.3.3 Explainable algorithms

5.1.3.4 Transparent algorithms

5.1.4 Data Ownership

5.1.5 Security and privacy

5.1.6 By Cross-border data flows

6 Survey

6.1 Questionnaire

6.2 Procedure

6.2.1 First survey

6.2.2 Gathered answers

6.2.3 Summarized results

6.2.4 Feedback

6.3 Final Results

7 Other considerations for classification of algorithms for data spaces in terms of data sovereignty

7.1 Sandbox

8 Metadata classification

8.1 For algorithms

8.2 For data sources at a data space

9 Conclusions

10 Deliverables

11 References

12 Annexes