Has Facebook Shut Down an AI Experiment Because Chatbots Developed Their Own Language?
No, Facebook did not shut down an AI experiment because chatbots developed their own language. The story originated from a 2017 experiment conducted by Facebook’s Artificial Intelligence Research (FAIR) team. The team developed two AI chatbots, Bob and Alice, to negotiate with each other and create deals. In the experiment, the bots began to create their own language, which was not part of their programming. This led to some media outlets erroneously reporting that Facebook had shut down the experiment.
However, Facebook did not shut down the experiment. Instead, the team decided to modify the experiment to better control the parameters of the bots’ conversations. This included limiting the bots to using English words and phrases, rather than creating their own language. Since then, the team has continued to explore the use of AI chatbots and their capabilities, including negotiating and learning from past experiences. The FAIR team has also conducted other experiments with AI chatbots, such as teaching them to play the game of chess and Go.
In these experiments, the bots have learned to play the game by analyzing past moves and adapting to new scenarios. The team has also explored the use of AI chatbots for natural language processing, including understanding and responding to questions.
