Over a year ago an anonymous Australian programmer started a social experiment called “Twitch Plays Pokémon”. The experiment consisted of a video stream on Twitch of the video game Pokémon Red. Viewers could interact with the video game by sending commands through the chat room of the stream which controlled the avatar of the game in real time. After more than 16 days of continuous playing, the game was completed.
When the experiment started, a few hundred people were watching. However, soon enough the internet caught on and the stream became unexpectedly popular. Instead of a few hundred, tens of thousands of people were watching.
During the remainder of the experiment, an average of 80,000 concurrent viewers was achieved, with a peak of 121,000 simultaneous viewers.
Within days, the experiment spawned many in-jokes, memes and even its own mythology around a specific item that was obtained during the play through. It became a culture of its own.
Throughout the experiment, around 10% of the viewers participated by sending commands through the chat room function. The result was complete anarchy and chaos. How was it possible to complete Pokémon Red in that lack of order?
Participants were all parsing on commands and in that lies the power of this experiment. The crowdsourced effort allowed participants to be heavy involved with the actions of the game. Many of the viewers were familiar with Pokémon Red and as such had a collective goal. They worked together as a team, as a flock, as a tribe. All these people had one thing in common, without even knowing each other. As such, even so-called “trolls” weren't capable of disrupting the play through for a long period of time. The group effort, the co-op play, of the stream was stronger than that.
Together, the participants created an autonomous entity. As individuals they sent commands, but together they formed a whole system. It is remarkable to see so many people working together towards a common goal. How could we implement this system in other ways, like solving societal problems? Or how can we implement this in intelligent systems, in which each unit behaves according to their own morale, but also according to a group morale? How can we network and connect individuals (be that human or artificial) to provoke behavior that is beneficial to the majority of these individuals?
When envisioning the relevance of these questions, it seems that there are many fields of interest that could benefit from this discussion: human-computer interaction, artificial intelligence and decentralized systems, but also societal development and perhaps even welfare and social work.
Subsequently, this raises a pertinent question: do we have the need, or do we need to join this debate?