BeeMe (MIT Media Lab)

Share This

BeeMe is a web-based social experiment that will be hosted by the Massachusetts Institute of Technology Media Lab at 11pm on October 31, 2018. The experiment, which takes place in the United States on Halloween night, allows internet users around the world to direct the actions of a live-action hero in order to defeat an artificial intelligence (AI) villain called Zookd.

Starting at 11PM ET, participants can log onto the MIT BeeMe website and control the hero’s actions by either submitting a command for the hero to follow or by voting a command up or down. The experiment’s slogan is: “See what I see. Hear what I hear. Control my actions. Take my will. Be me.” All participants in the experiment will see and hear events from the hero’s first-person point of view (FPV).

The objective of this crowdsourcing experiment is to gather data about how smoothly a large group (the hive) can make an individual carry out a series of tasks. The challenge will be for participants to work cooperatively; it is expected that when the group is successful, the hero’s actions will be fluid, but when the group fails, the hero ‘s actions will be sporadic, random and perhaps even zombie-like. Commands that might endanger the privacy, dignity or safety of the human actor who is playing the hero will not be allowed.

The MIT Media Lab has a tradition of providing internet users with innovative, interactive Halloween treats. In 2016, they released Nightmare Machine, a project that used deep-learning algorithms to generate scary zombie portraits. Website visitors were then asked to help the Nightmare Machine’s artificial intelligence learn by voting on whether an individual portrait was ‘scary’ or ‘not scary.’ The project also used deep learning algorithms to determine what visual characteristics make a house look haunted or a city look toxic and then applied what they learned to famous landmarks around the world. In 2017, the MIT Media Lab released Shelley, a collaborative writing project that paired humans with deep-learning AI to create a horror story anthology.