The OECD and the Global Partnership on AI have published a key report on the openness of AI systems – and Croatia actively participated in its development! Juraj Bilić from CARNET, project leader of the BrAIn project, made a significant contribution as a member of the OECD Working Group on Artificial Intelligence Governance.
The report answers key questions: How to balance innovation and security? When is openness in AI models beneficial and when risky? What does this mean for digital sovereignty and talent development?
U izvješću se posebno ističe:
- open models encourage innovation, competition and talent development, increase transparency and enable digital sovereignty
- Open models simultaneously open up space for abuse, privacy risks, uncontrolled spread, and exploitation of vulnerabilities.
The OECD suggests a marginal analysis approach: instead of asking whether openness is good or bad, it is important to consider what additional risks or benefits arise from opening up the model in a specific context. The goal is not openness for openness's sake nor closure for security's sake, but the development of reliable AI that benefits society as a whole.
CARNET's BrAIn project has also been highlighted as an example of good practice in introducing AI into education. Find out how Croatia is shaping the future of artificial intelligence!
