Facebook's Artificial Intelligence bots create own language

Facebook's Artificial Intelligence bots create own language

Mark Zuckerberg may believe Elon Musk's warnings about the dangers of artificial intelligence are "irresponsible", but a recent incident at the social network appears to suggest the Tesla boss could have a point: Facebook researchers chose to shut down an AI they invented after it started speaking its own made up language.

However, researchers at the Facebook AI Research Lab (FAIR) found that they had deviated from script and were inventing new phrases without any human input.

There has been a real flurry of interest in the last couple of days in a couple of chatbots (reassuringly named "Bob" and "Alice") developed by Facebook AI Research.

"There remains much potential for future work, particularly in exploring other reasoning strategies, and in improving the diversity of utterances without diverging from human language", they concluded.

There isn't enough evidence to prove that such unexpected AI divergences could be threat machines eventually taking over their human operators, but it will certainly make development more hard as people are unable to grasp the overwhelming logical nature of languages developed by AI.

The issues with the chatbots arose as a result of them independently discovering that using English as a means of communication was ineffective, with them creating their own language as a result. Initially, the agents were communicating in pure English but later began to drift towards using their own language.

Sentences like "Balls have zero to me to me to me to me to me to me to me to me to" were unintelligible to him and his peers, but to the conversing AIs they were a shortcut to speedier understanding in the negotiation process.

The Facebook Artificial Intelligence Researchers (FAIR) set up an experiment where a number of chatbots were trained to become adept negotiators. But unfortunately, Facebook made a decision to pull the plug.

Bot 2: "Balls have zero to me to me to me to me to me to me to me to me to". The robots had the task of negotiating to exchange various virtual objects. "They would, for instance, pretend to be very interested in one specific item so that they could later pretend they were making a big sacrifice in giving it up", said a paper published by the Facebook Artificial Intelligence Research Division.

After detecting this new language, researchers had to reprogram the artificial intelligence to prevent it from changing the language codes.

The bots were created as part of a program to teach computers and machines how to negotiate, while the ultimate goal was to develop personalized "digital assistants" who would be able to communicate with people.

The conversation continues with variations of the number of the times Bob said "i" and the number of times Alice said "to me" in the discussion.

Unfortunately, though, it was too efficient for humans to understand.

Related:

Comments

Other news