In recent years, privacy and climate have emerged as two important themes in European policy – two examples of this are the General Data Protection Regulation (GDPR) and the European Green Deal. Nevertheless, the two themes appear to be diametrically opposed to each other in some respects, says Ana Oprescu, one of the research leaders. ‘Sometimes it seems as if the GDPR and the Green Deal are not part of the same universe. Through this project, we are trying to bring them closer to each other.’

Energy-hungry technology

The research project focuses on the question of which technology aimed at protecting privacy-sensitive data can best be used and in which situation. There are various privacy-enhancing technologies, but some of them consume a lot more energy than others. One frequently used privacy-enhancing approach is replacing real data with synthetic private data, in which an algorithm generates statistically similar ‘fake’ data that can no longer be traced to specific persons. Because synthetic data largely has the same characteristics as the real dataset, the data can still be used effectively for all kinds of research purposes.

Synthetic data is a promising technology that is used more and more often by companies, says Oprescu. ‘Synthetic data safeguards privacy, while hardly any information is lost. In that respect, it works better than other technologies, such as k-anonymity, where information is omitted for the benefit of privacy and thus the data utility might be reduced. However, there is one major disadvantage to synthetic data: it is an energy-hungry technology, which therefore has a relatively bigger impact on the climate.’

Every situation is different

The researchers are considering how three important data values – privacy, utility and sustainability – are related to each other precisely, and how they can best be balanced. They are going to develop a decision-making framework on the basis of which it will be possible to examine which technology is most desirable on a case-by-case basis.

‘Which technology is optimal depends on various factors’, says Oprescu. ‘What do you want to do with the data exactly, which information can you obtain from it and which degree of privacy do you need? That differs per situation. In some cases, for example, privacy may not be so important that the use of energy-hungry synthetic data is justified, while in privacy-sensitive situations you can’t do without synthetic data. When you have a complicated research question, such as analysing how women are underrepresented among senior university staff members, you are faced with very personal data and relatively small numbers. In that case, you really need advanced technologies, such as synthetic data, because too much relevant data is lost with technologies that consume less energy.’

Collaboration with partners from industry

Researchers from various fields are working together in the project: computing science, economics and law. In addition, the researchers collaborate with partners from industry – BlueGen.ai, KLM and GVB – to learn how these partners deal with privacy, data utility and energy consumption. That interdisciplinary approach is important, explains Oprescu. ‘It is best to examine such a complex issue from different fields. It is not only about the technological side of the problem, but also about values, legal frameworks and human behaviour. With our project, we want to stimulate dialogue between the business community, management and software developers, because we can only find the right solutions by looking at the problem together.’

The intersection of digitalization and sustainability in Europe will be further explored at the ACES conference Sustainable Digitalisation in Europe (21-22 March), which is organised by the researchers involved in this project.