Let machines unlearn – machine unlearning and the right to be forgotten

Abstract

The era of machine learning has prospered quickly although it has rather a short history. Massive customization, which was considered as costly task, is now a concept applicable concept to almost all systems, and individuals gain the advantage of convenience from such technologies. However, disadvantages of machine learning does exist. Specifically, the management of privacy data and handling of deletion requests of such data are the challenges facing machine learning. Consequently, a way to unlearn what has been learnt by machines has been introduced: machine unlearning. Regarding the deletion of privacy data, the right to be forgotten is the right to support one’s informational autonomy by giving the decisive power to data providers. Though suggested recently, the importance of the right becomes greater, especially in European countries. In this emergent research forum study, we investigate the possible interplays between the right to be forgotten and machine unlearning, and explore future research directions.