Many of us had tried creating our own secret language as kids, and apparently, Facebook’s AI chatbots were inspired to do the same thing. Facebook researchers recently found a couple of chatbots named Alice and Bob communicating with each other in a language they had generated by themselves which a human will be unable to understand.
The conversation between Bob and Alive went like:
Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to
Bob: you i everything else . . . . . . . . . . . . . .
Alice: balls have a ball to me to me to me to me to me to me to me
Bob: i i can i i i everything else . . . . . . . . . . . . . .
Alice: balls have a ball to me to me to me to me to me to me to me
Bob: i . . . . . . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to
Bob: you i i i i i everything else . . . . . . . . . . . . . .
Alice: balls have 0 to me to me to me to me to me to me to me to me to
Bob: you i i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to
It appears that it is not actually a language but some gibberish sprouting due to an error but researchers say it is a kind of shorthand. A visiting researcher at Facebook Artificial Intelligence Research division, Dhruv Batra said, “Agents will drift off understandable language and invent code words for themselves. Like if I say ‘the’ five times, you interpret that to mean I want five copies of this item. This isn’t so different from the way communities of humans create shorthands.”
Facebook recently announced their initiative for developing AI bots that were able to negotiate at Facebook Artificial Intelligence Research (FAIR). The same bots were found to have developed an ability to lie, and creating a secret language is their latest triumph.
The software bots are meant to communicate with both humans and other computers and perform tasks like making recommendations and booking appointments. Facebook researchers left the bots to communicate among themselves as they liked, and they did not even have to limit themselves to English. The bots then began to deviate from the script to improve their deal-making efficiency.
We won’t be getting comfortable with AI inventing its own stuff for a very long time, but the Facebook chatbot secret language is not the first kind of its happening. Alphabet’s DeepMind has created an AI named AlphaGo that works on the same principle as Facebook’s AI bots. Companies like Amazon and Google are making massive investments in machine learning and artificial intelligence. Amazon Alexa and Google’s digital assistant, as a result, are moving towards better finesse that will soon be able to perform tasks that we may have never imagined in the past.
Facebook has since killed the chatbots stopping them from creating secret languages only because that was not their intended purpose. The incident leaves us with many questions. Would we ever be comfortable with AI becoming this intelligent? Is there a need for researchers to put an end to this? How useful is the artificial intelligence of this kind? What kind of check and balance is required before this turns into a sort of a disaster?
