Facebook AI evolved, humans kill it


What really happened and should we be worried

So by now you have no doubt heard about the Facebook AI that was created by Facebook’s Research team. The AI was tasked with solving problems and incentivized to do it quicker as it learned.

However the AI or two AI’s as it was who were speaking to each other quickly discovered that English was not the most efficient way for them to solve the problems set before them. So what did they do, well they made up their own language based off of English, an almost short hand style.

This is fantastic in many ways as it shows that machines can learn, even beyond the basics of what their human programmers programmed them to do, albeit in a simple form this time.

Facebook decided that the AI was not really doing what they intended so they shut it down. Effectively killing it, however the AI was not self aware and was just a code process. None the less it goes to show what might happen if AI as is being discussed was put to task on public jobs, to make important decisions. What if we used it in an uncontrolled environment where it suddenly decides that it know best, cuts our human access to it and then does what it likes. This is the problem Stephen Hawking and others have spoken about.

View more  Engie App Review

What if we create unknowingly a Skynet style AI some day. For now there is no need to worry, but once the people in power decide they want AI running things like military installations or nuclear plants, then worry.

Read more from Facebook’s own researchers below:


Facebook AI evolved, humans kill it was originally published in Ireland’s Technology Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.