The OECD report released on February 6, 2025, points to the urgent shortage of quality data facing AI developers, a problem for which the OECD's recommendation on data access and sharing offers a solution. The document highlights that improving data sharing and access could represent an economic advantage equivalent to 1-2.5% of GDP in OECD countries.
According to the report, 29% of large companies used artificial intelligence in 2023, while this ratio was only 8% for small and medium-sized enterprises. Three levels of data sharing were defined: conditional access, non-discriminatory paid access, and completely open data sharing. The document emphasizes that AI models must have FAIR principles-compliant properties: findable, accessible, interoperable, and reusable.
The 2025 recommendation pays special attention to the role of privacy technologies in facilitating secure data sharing. It proposes two key solutions: privacy-enhancing technologies (PETs) - including synthetic data generation and secure execution environments - and trusted data intermediaries (TDIs). While PET technologies enable the secure sharing of sensitive data, TDIs provide institutional guarantees for reliable data handling, ensuring a balance between data protection and innovation-supporting data sharing.
Sources:
1.
Enhancing Access to and Sharing of Data in the Age of Artificial Intelligence
This OECD report discusses how governments can enhance data access and sharing to maximize the benefits of AI while addressing related challenges.
2.
