With the drive for AI (artificial intelligence) devices such as Siri, Cortana and the like, not stopping anytime soon, Facebook invented their own with amazing results. When given their task and no rewards or incentives to stick to English, AI robots nicknamed Alice and Bob, circumvented the Language and instead invented their own. Shortly there after, the experiment was stopped as scientists marveled at the results. How could a machine programmed to use English deviate from it and invent their own language?
The experiment
Wanting to better understand how AI works and the multitude of different ways it could do so, Facebook invented their own that was tasked with trading and bartering for simple items.
Their instructions were to determine the value of each hat, book, and ball, then trade and barter for it. With no motivation, incentive or reward to sticking to English for communication, the robots quickly took an unexpected turn and invented their own undecipherable language. According to Facebook researcher Dhruv Batra, this should have been expected as machines have been known to deviate from intelligible words and create their own form of short hand as crowds of people do. In other words, AI robots are becoming ever more human like. Before people realize it, we just might have C-3PO and R2-D2 type AI robots.
The results
After having watched this momentous occasion, researchers were concerned and baffled, but also astonished at the robots' capabilities.
On the one hand, if they were able to invent their own means of communication on the spot, what is next for them? On the other hand, they were able to successfully complete the given task. On its own, the latter is a huge accomplishment. AI robots-machines able to not only recognize the value of a given object but also barter and trade for others is a huge step.
It shows that these machines are able to not only comprehend ever more difficult tasks but also learn and adapt as need demands.
By the end of the decade, we just might have AI so advanced, that it will be near impossible to tell the difference between it and an actual person, save for outward appearances. Already researchers have said that when people communicated with their artificial intelligence robots, it was extremely difficult for them to tell the difference between it and an actual person.
While this can be good for tasks such as a personal assistant, it can also lead to irritating results such as it being impossible to understand and translate the AI language. As such, it could be a bit of a problem when it comes to reprogramming and updating. Here is hoping the robot uprising does not come to pass anytime soon.
