Facebook Did Not Shut Down Chatbots Out of Panic
The Claim
Two AI robots that Facebook created had to be shut down because they started creating their own new language.
News posted on
Emerging story
A number of recent posts have claimed that Facebook needed to shut down two AI robots it had created because they were communicating with one another in their own language. These posts are often accompanied by pictures of scary-looking robots, implying that these robots were sinister or some kind of threat. These posts have gained hundreds of interactions on social media.
Additionally, a number of news updates upon this story claimed that the Facebook engineers panicked after the robots developed their own language, which led to the shutdown.
Misbar’s Analysis
Facebook did not shut down two AI bots that created a language of their own. Instead, they redirected them to speak English. Further details within this story reveal that it was not at all a suspicious or sinister occurrence.
As a number of articles have highlighted, Facebook employees did not panic due to the bots creating their own language. Instead, the programmers made an error in not instructing the robots to maintain English as their language. They needed to reprogram them because, as a Facebook research scientist told Fast Company, “Our interest was having bots who could talk to people.”
It is not unusual for AI bots to create their own language to boost efficiency. In fact, Google Translate uses its own self-created AI language when translating one human language to another.
Another element that makes this story less “sci-fi” than the social media posts suggest is that the AI bots were not physical robots that had the appearance of androids but computer programs that were used for a department called Facebook Artificial Intelligence Research.
Some of the posts also claimed that this research happened recently. However, this story dates to 2017.
