BeeMe was an 📝MIT Media Lab social experiment staged on Halloween 2018 in which internet users collectively controlled a consenting actor's every action in real time, streamed live.
The experiment began at 11 p.m. ET on October 31, 2018 and was built around a fictional premise: an evil artificial intelligence named Zookd had been accidentally released online, and the audience had to coordinate at scale to help the actor stop it. Participants submitted instructions through a web interface and voted on which ones the actor would carry out. Anything illegal, or anything that violated the actor's safety or identity, was prohibited; beyond those limits, the crowd's collective will governed what the actor did.
BeeMe was led by Niccolò Pescetelli, a researcher studying collective intelligence, within Iyad Rahwan's Scalable Cooperation group at the Media Lab. Its research interest was coordination — whether a large, anonymous crowd could converge on shared intent rather than fragment into noise. More than a thousand people logged on. Technical problems with login and voting pushed organizers to fall back on chat, and while many participants submitted requests unrelated to the plot, the crowd ultimately steered the actor to stop Zookd.
The project belongs to the lineage of live 📝Interactive Media, where audience input drives the narrative rather than merely responding to it. Press coverage ahead of the event repeatedly likened the premise to an episode of 📝Black Mirror.