Artificial Intelligence: Facebook neutralizes its robots who’ve developed their own language


Image via Pixabay

Two conversational Facebook robots controlled by artificial intelligence (AI) had to be reprogrammed after they had started exchanging in a language that researchers responsible for the study could not understand.

The conversation between the two robots took place in the context of an experiment carried out in June 2017 by researchers affiliated with Facebook. The robots had the task of negotiating to exchange various virtual objects.

After a few rounds of negotiations, however, the robot team noticed that artificial intelligence had invented a new language and that robots used it to conclude contracts.

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 me

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

Bob: i . . . . . . . . . . . . . . . . . . .

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

Facebook’s research on artificial intelligence should be used to create a personal assistant capable of imitating human behavior.

This article has been viewed 1915 times. Article originally posted: August 6, 2017, 7:57 am (UTC-0). Last update: August 6, 2017 at 7:57 am (UTC-0).

Facebook Comments