The team at the Facebook #Artificial Intelligence Researchers (FAIR) had the surprise of their lives after two of their AI #Bots communicated in a language totally incomprehensible to the human mind. According to The Independent, a pair of AI was given instructions to negotiate a trade that involved hats, balls, and books. As with all kinds of trading, the two artificial intelligences, named Bob and Alice in the transcripts, were instructed to find ways to negotiate with one another with the goal of improving their communications.

The conversations became weird quickly when the two bots started speaking in their own version of English.

Advertisements
Advertisements

Apparently, the report said, the two artificial intelligences improvised on the #english language so they can communicate better. However, Techly reported, the result was totally bizarre and unintelligible to the researchers. At one point, Alice said “balls have a ball to me to me to me to me to me to me to me” to which Bob responded with: “i i can i i i everything else . . . . . . . . . . . . . “

AIs creating new language only natural, researcher says

The negotiations were part of the FAIR experiment that began last month. According to the FAIR Team, they intended to teach the bots to negotiate in order for them to develop “complex communication and reasoning skills.” The first part of the experiment forced the AIs to conform to understandable language, but the second part gave them the freedom to deviate and invent their own code of communication.

Advertisements

FAIR Researcher Dhruv Batra explained to Fast Co Design that it is natural for the AI agents to invent their own language in order to communicate better. The researcher likened this to how humans create shorthands to convey information across to one another. The researcher also noted that there was “no reward” to forcing the agents to stick to English.

Mark Zuckerberg and Elon Mask Clash Over AI

The topic of artificial intelligence is a cause of debate between tech giants Elon Musk and Mark Zuckerberg, Facebook’s chief executive officer. In July, The Guardian reported, Musk reiterated his call for regulation regarding AIs, calling them a “fundamental risk to the existence of civilization.” Musk spoke before U.S. Governors last month to warn them about the perceived threat of AI and the importance of regulation.

In response, Zuckerberg called during a Facebook Live Broadcast what he described as “doomsday scenarios” as negative and “irresponsible.” The Facebook CEO added that he is upbeat about his company’s progress on AI research. On detractors, Zuckerberg said he doesn’t understand those who make up the end of the world scenarios about AI.

Advertisements

“It’s really negative and in some ways, I actually think it is pretty irresponsible,” Zuckerberg said in response to a question asking him for his thoughts on AI and the world’s future. In addition to AI, Zuckerberg is also interested in virtual reality. In April of this year, the Facebook CEO explained his vision of VR during the F8 Conference in Silicon Valley.