“The police recognizes that AI will have an unprecedented impact on our society and wants to be at the forefront of AI developments. Together with partners in academia and business, the police can employ state-of-the-art AI techniques to improve the safety in the Netherlands in a socially, legally and ethically responsible manner.”
What kind of research within AI is the Police Lab AI working on and why?
“In the Police Lab AI, Utrecht University (UU) and the University of Amsterdam (UvA) are working on techniques across the full breadth of AI. The UvA excels at developing new techniques for extracting important information from different sources such as photos, text and video. The strengths of the UU are complementary: we develop formal models for reasoning with and about the information that is found in the data. In these models, aspects such as transparency, privacy and explainability are just as important as algorithmic aspects such as accuracy, computability, efficiency, effectiveness, and formal correctness.”
What will surprise us the most in AI over the next five years?
“That it will become commonplace and accepted as a simply a technology that aids us in our daily affairs. As John McCarthy, one of the founders of the AI discipline said: “As soon as it works, no one calls it AI any more.””
What is the biggest misconception about AI?
“That AI is only data-driven: give data to a computer and it learns all by itself. Humans invented the games of chess, Go and Jeopardy, and the computer needs explicit models of these games in order to function. The models underlying police processes such as intake and investigation are different from these ‘simple’ models of board or TV games. No amount of data will teach the computer to talk to civilians and do crime investigation all by itself – we humans have to teach those things to the computer.”
Floris Bex is scientific director of the Police Lab AI, Professor of Data Science and the Judiciary at Tilburg University and associate professor at Universiteit Utrecht.