The MIT Lab has an excitingly creepy way for you to celebrate Halloween, if parties or candy aren’t your speed. Next week, you’ll be able to work with other internet users to control one person’s actions for an evening, as spotted by Business Insider.
Researchers at the lab made a program that lets a group of users control a human being while hearing and seeing everything the person does as if they were inside the person’s mind. The human is an actor hired by MIT Media Lab. The project is called BeeMe, and it’s billed as a web-based social experiment.
Starting on Halloween at 11PM ET, people can log onto the BeeMe website and start suggesting and then voting on what action the actor should take next, like opening a door or entering a room. (Naturally, commands that would endanger the well-being, privacy, or dignity of the actor will not be allowed.)
As the outlandish narrative the MIT researchers have crafted goes, the human being has given up his free will and agency over to a group of humans in order to defeat an evil artificial intelligence called Zookd that has been accidentally released online.
The idea behind the social experiment seems to be partly inspired by “White Christmas,” an episode of Black Mirror where a dating guru played by Jon Hamm and several of his clients watch and give advice to a man through a chip embedded in his head as he approaches women at a Christmas party.
The implications of this are dystopian for a number of reasons. In the episode, the guru and his clients are able to see what the man sees and hear what he hears, part of what makes it such a violation of privacy when the man ends up going home with a woman from the party. Ultimately, the guru gets caught by law enforcement, and his voyeuristic activities get him labeled a sex offender.
The MIT Media Lab’s experiment is understandably more light-hearted and data-oriented. The researchers hope to garner more data on whether a group of people are capable of making one person execute a fluid series of tasks, or whether the results will be disjointed and random. Think “Twitch Plays Pokémon,” but with a real player.
MIT has a history of conducting creepy experiments around Halloween. In 2016, the Media Lab created an algorithm called Nightmare Machine that makes images spookier. Participants clicked on scary images to teach the AI what qualified as truly grotesque, in hopes that it would learn as time went on.