Collaboration centers around the regulatory sandboxes are recommended as the countries will be able to share experiences on how these can be organised most innovatively and efficiently. Initiatives on establishing such a collaboration center should be launched in the near future, as setting national sandboxes will be stipulated by the AI Act. The Norwegian Data Protection Authority might be an interesting and relevant partner as the institution holds considerable experience in sandboxes tailored for responsible AI. | |
The Nordic and Baltic countries differ from the majority of the EU member states as the Nordic-Baltic countries´ public sectors rank among the most digitalised in the world, and almost all public service and exercise of public authority depend on – assumedly in the light of worsened treats – too vulnerable systems. Therefore, EU information and cybersecurity regulations are inadequate to ensure robust public services and administrations. As information and cybersecurity should not be regarded as issues only relevant to the intelligence services, promoting robustness via coordinated regulatory initiatives in the Nordic-Baltic counties is recommended. | |
Data sharing, hereunder establishing more extensive sector-specific databases for developing machine learning-based models and other AI systems, can become instrumental in ensuring human rights, e.g. reducing biases against protected groups. Such cross-border collaborations require, however, a stronger focus on harmonisation, particularly in the regulatory terminology, semantics, and data definitions. Further, the researchers have debated whether a Nordic-Baltic Council of Data Ethics is recommendable. Here, pros and cons have led to the DigiLaw project recommending a debate hereon. |