Researchers at FAIR had developed two chatbots that were intended to replicate the way humans barter and trade. The two bots were to trade balls, hats, and books, assigning a value to the objects, and then bartering those objects between themselves. Using machine learning algorithms, the chatbots, also known as dialog agents, were also programmed to utilise human speech in the bartering process as part of the experiment. However, it soon became clear that the chatbots had deviated from normal speech, and instead, the artificial intelligence software driving the experiment, had invented a new language. The part that has caused the most alarm to some, is that this was done without input from the FAIR researchers. Here’s a look at the conversation that the chatbots created below:
This was enough to send a good number of people into a state of panic. The fear appears to be driven by the notion that, a sentient artificial intelligence system could be a serious threat to humans. Even Elon Musk has recognised the potential threat that an advanced artificial intelligence system can pose to mankind. Musk is so concerned of the dangers of artificial intelligence, that he is a regular advocate for the imposition of strong regulation on the artificial intelligence space. Now despite all of this, it is clear that the research conducted by FAIR is unlikely to produce any human enslaving robots any time soon.
As Dhruv Batra, who is an artificial intelligence researcher at Facebook explained,”Artificial intelligence agents inventing their own language is a well-established sub-field of artificial intelligence research. What took place between the chatbots was nothing out of the ordinary in the field, so there really is no need for panic. The language that the chatbots created can almost be seen as a form of short hand.”
“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 shorthand.”
So, whilst it is clear that the wider community most definitely overreacted to the FAIR chatbot experiment, the debacle has raised a wider issue that many appear to be frightened of. Is the development of artificial intelligence necessarily a good thing for mankind? No one is certain where the development of artificial intelligence research will take us, and for some, that is a scary prospect to entertain. The question that looms is whethere we will have to ensure that artificial intelligence systems are well regulated to prevent some sort of a Skynet-esque apocalypse? It is only with the passage of time that the answers to these questions will really be known to us.