Will AI make privacy guidelines easily accessible?

ML2Grow will find out through an experiment.

Informing citizens and professionals about the processing of personal data is one of the tasks of the Data Protection Authority (DPA). It does this via its website, where all information is made available in the form of recommendations and decisions. Because of the multitude of information, which is still growing, it is not always easy for the user to quickly find the right information. For this reason, the GBA is investigating how artificial intelligence (A.I.) can help the user to obtain the correct information online more quickly and efficiently.

Specifically, the GBA wants to expand the possibilities of Conversational A.I. test using a prototype. Such Conversational A.I. systems ensure that users can have simulated conversations with a computer using Machine Learning and Natural Language Processing. Because the A.I. is so smart, it looks like chatting with a human. This is reminiscent of the well-known chatbots, yet the Conversational A.I goes a step further. A simple chatbot is based on a fixed question structure and can therefore not answer complex questions that are not included in the question structure. Conversational AI can do this. The system makes connections between topics, which gives the conversation a big boost.

The experiment will be started in October 2021 in collaboration with ML2GROW, a spin-off of Ghent University (UGent) and imec. Nido, the government innovation lab at FPS BOSA, supports the GBA in this challenge. In addition to security, data protection and reliability, speed and relevance will be key metrics in the experiment that will start at this early stage of development with a limited group of users.

Gilles Deweerdt

Newsletter

Sign up for our newsletter.