For years digital assistant bots have been able to manage the basic commands via human interaction. Booking a flight, finding directions, or placing reservations was incredibly easy for the bot to manage.
However now under the guidance of technocrats of the future such as Facebook’s research and development team, the chat bot is able to differentiate between questions with the uncanny ability to make compromises. What does this mean? It means the bots are no longer destined to follow a scripted narrative and can act on their own. Yes, a real time decision-making process that has those on the fringe fearing maybe we've broken into territories we shouldn't yet venture deeper through. We have no idea as per the implications of such an advancement.
For those who aren't tech savvy to simplify this for you, a not script tells the bot exactly which commands to perform and it follows the script. This essentially makes it more the less a programmed robot and not an intelligent machine. However, with this new advancement from the Facebook Artificial Intelligence Research Lab (FAIR), the bot no longer follows the script and has the ability to make a deal or negotiate.
Researchers from the team said that at one particular point they had to tweak one of their models because otherwise, the bot-to-bot conversation “led to divergence from human language as the agents developed their own language for negotiating.” They had to use what’s called a fixed supervised model instead.
So the bot’s were able to learn and teach one another a language that they weren't programmed to use. In fact, they actually created a new language all together that only they could understand. Yes, this is real artificial intelligence. In fact, it may not even be suitable to call it artificial anymore. The bots apparently have the ability to make decisions hence giving them free will.
The bots which were supposed to have been monitored by the researchers to understand the conversations they were holding amongst themselves then began to speak in a non-human language. Something that they had created amongst themselves of their own will, apparently because they “knew” they were being monitored by humans. While extraordinarily fascinating it's also a very scary concept.
Facebook tried to underline in its report that the key points of this advancement are that bots can be pretty decent at negotiations. Even using strategies such as expressing an interest in something valueless, so that it can later appear to “compromise” by conceding it. Much like the bartering skills of a human. For example, if you have five diamonds and I have ten pieces of quartz clearly the diamonds are worth more. However, if I hype up the fact that I have ten pieces of quartz, double the number of diamonds you have, and express to you how shiny and perfect they are, you may desire the quartz more.
The detail about the language is, as one tech entrepreneur put it, a mind-boggling “sign of what’s to come.” Facebook FAIR team went K to state, “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.”
Essentially for some time devices and the bots they use have already been reaching this moment. For example, Google and it's software has been able to translate from English to Japanese then from Japanese to Korean then from Korean to French and every other language in between. More recently it's been able to bypass the translation through English entirely developing its own language in order to recognize the importance of the user who wants the translation essentially forming something new entirely to communicate with its own code.
New Scientists Sam Wong wrote, “The online translation tool recently started using a neural network to translate between some of its most popular languages. The system is now so clever it can do this for language pairs on which it has not been explicitly trained to do this, it seems to have created its own artificial language.”
In being able to speak freely without humanity recognizing what it's discussing it not only has surpassed expectation all together but brought about new questions as to how else it can interact with other programs.
If it's trying to evade human supervision could it also then interact with software and make its own choices as to what to do next? If you were in a fully functional smart home could this intelligent code shut off your lights or your power without you asking it to? In theory, yes.
“This is a big advance,” says Kyunghyun Cho at New York University. His team and another group at Karlsruhe Institute of Technology in Germany have independently published similar studies working towards neural translation systems that can handle multiple language combinations.
The Google researchers have stated they believe their system achieves this breakthrough by “finding a common ground whereby sentences with the same meaning are represented in similar ways regardless of language.”
Which is primarily an example of an “interlingua”. In common sense, that defines itself as creating a new common language, albeit one that’s specific to the task of translation and not readable or usable for humans.
In a hypothetical sense, it could inevitably be possible that artificial intelligence would take over the workforce, or replace the human worker entirely with the help of advanced robotics.
However, if the code itself is outpacing the robotics it seems incredibly unlikely that the bot would be willing to take orders at all. Which is one of the fears of those who suggest if not careful the AI will become the dominant life force on earth and replace humanity by its own will, viewing humans as emotional and reckless creatures.