Artificial Intelligence in Data Sharing
Without data, no machine learning is possible. The more available relevant data, the better the predictions, thus improving the usefulness of the machine learning applications. Access to data is, therefore, crucial. In the Netherlands, data is often kept locked, usually for legal or commercial reasons. To break through these existing barriers, sharing data must be organised efficiently and responsibly, but much faster and more efficiently than we are used to at the moment. Increased access to data means an acceleration of AI implementation and higher accuracy, which results in overall better service. Trust, insight, more knowledge, and our democratic principles form the basis.
Any organisation must always carefully weigh and categorise opportunities and risks to protect business interests and ensure that their data is not misused. The question is: how are we going to feed AI applications with data with these restrictions in place?
The Netherlands AI Coalition wants to simplify data sharing, a crucial element of AI, by researching, further developing, and deploying privacy-enhancing techniques (PETs). Knowledge about what is and is not allowed should be readily available. Besides, ecosystems around a specific sector or region must be encouraged to consider data sharing solutions.
The Data Sharing working group provides participants with knowledge and resources around responsible data sharing. There are also training courses for different target groups. An extensive report is readily available containing detailed information about data sharing in AI, the possibilities, the preconditions, and which processes can be followed. The knowledge base is put into practice through use cases. These use cases are identified per sector and carried out by the member organisations.