In the future, games will be able to understand the player and make decisions; yet the technology could revolutionise many other areas. Here, PROFESSOR GEORGIOS N. YANNAKAKIS explains how, and how the same game artificial intelligence methods could be transferred to the classroom, hospitals and beyond.
Ever since videogames were invented, users have had to adapt to them: We have to work with the difficulty levels pre-determined within the game and we have to follow the route pre-designed for the game. Yet the future of gaming is a wholly different universe according to Professor Georgios N. Yannakakis, an associate professor within the Institute of Digital Games at the University of Malta (UoM) and author of the first comprehensive textbook on artificial intelligence (AI) and games.
“Depending on the sensors your console or PC would have, and based on what the software itself could process, games in the near future will be able to read anything from facial expressions to behaviour and physiology (including heartrate) to adapt to the player,” he reveals.
As things stand, players currently learn a lot about the game at hand but the games do not… But new applications using AI could change that, leading games to adapt to the user to minimise frustration and maximise engagement, among other things.
“Once a game knows something about the player, then it can change accordingly,” adds Professor Yannakakis, whose seminal paper written along with John Hallam about a modified version of the iconic Pac-Man arcade game (in which the ghosts change their behaviour according to the player’s strategy), was picked up by various international media outlets, including The Guardian.
“It can create new emotional states, aesthetics, sounds, or even levels… Moreover, player-adaptive games could even change the storyline depending on the choices players make, which could include anything from endings to dialogues.”
All this could lead to an incredibly rich experience for the player, but there are also other questions researchers and developers must ask themselves and find an answer to: “Games are currently very controlled environments that demand very particular cognitive skills from the player… This is what users are accustomed to. So, to which degree do you adapt and how long do you adapt the game for? And do humans want to lose control of the game?
“Then we also need to think about the other uses AI could have, such as in automated car driving systems, for example. Would people be okay with this? And to which degree should the system intervene?”
As Sophia, the first robot to ever be given citizenship, takes the world by storm, the rise of AI in everyday things seems unstoppable: From health-related applications used in hospitals to educational tools for children and even for simulation purposes, the future, as it goes, belongs to AI.
“One EU funded project we were recently working on was on a game that could detect what type of dyslexia a user has, making it easier for educators to deal with the situation as hand,” Professor Yannakakis says. “Moreover, I am also working – as part of a team – on Envisage, a H2020 project about transferring the mature technology of game analytics to classroom analytics, meaning that behavioural patterns of students would be analysed and visualised to help teachers teach better.”
With a myriad of projects in the pipeline – including eCrisis, an Erasmus+-funded project to develop games that help foster debate between students and teachers about refugees, the far-right and social exclusion – AI is as hot a topic at the University of Malta as it is anywhere else in the world. So, while we don’t know what the future holds, as things stand, Malta could play an integral part in developing the next generation of AI.
Help us fund more projects like this, as well as research in all the faculties, by donating to RIDT. Click here for more information on how to donate.