Artificial Intelligence is not out there in the utopian future anymore. It is already here, and it is changing the everyday interactions with the objects and services we use every day. Thus, today’s designers have to be familiar with the additional challenges and complexities the medium of machine learning adds to the current design process, as well as be capable to prototype such intelligent objects or services.

Entertainment is already one of such areas that is being touched by machine intelligence. Our team has set ourselves to discover new possibilities in that field, provided with the capability to prototype with computer vision and pattern recognition.

Doodle Battle is a board game for up to four players, encouraging competing with the machine in terms of doodle recognition. The rules are simple: the machine tells the players a clue, which players have very little time to decipher and draw a doodle of the described object on a post-it. In order to gain a point, the player has to be the first to draw a doodle and put it on a designated area on the table. If the machine recognizes the doodle correctly, and if the recognized object is the correct one, the player wins the round.

However, just as humans are not perfect, trained machines are not perfect. Machine learning algorithms quite often misclassify the doodled objects, adding the sense of imperfection and making the game even more interesting in terms of human-computer interaction.