Zuckerberg V/S Musk- What is wrong about AI?

Courtesy: RiversideEddy

Researchers at Facebook Artificial Intelligence Research (FAIR) were forced to shut down a pair of chatbots after they discovered that it developed a secret language of it's own. Here's the excerpt:

Facebook's AI language

Bob: i can i i everything else . . . . . . . . . . . . .

Alice: balls have zero to me to me to me to me to me to me to me to me to me

Bob: you i everything else . . . . . . . . . . . .

Alice: balls have a ball to me to me to me to me to me to me to me

Bob: i i can i i i everything else . . . . . . . . . . . .

Alice: balls have a ball to me to me to me to me to me to

Bob: i . . . . . . . . . . . . . . . . . . .

 This has added fuel to the verbal spat that Musk had with Zuckerberg. Musk has his longstanding concerns regarding the apocalyptic impact of AI on humanity while Zuckerberg has been somewhat optimistic as he is about any of his other projects saying that "naysayers" with their "doomsday" concerns are pretty negative" and "irresponsible." This was an indirect hit at Musk who took to Twitter to say that Mark's understanding on this subject is "pretty limited."

This brings us to the question- What's next? 

Let's take a step back and analyse what AI is. AI is a self-learning intelligent machine like we are. AI has become an overused term that it is used for pretty much everything and how it is portrayed in science fiction makes us think of only the extremes- utopia or dystopia. It is not just that. The concerns are perfectly alright but AI is not all bad or all good.

Let us analyze the scenario of Facebook. If we think that FAIR shut down the robots because of an existential threat on humanity, NO! It was because it found something that should not have been there and was not even the subject of interest. Neural Networks on which AI chatbots are built is difficult to understand and we still do not completely know how it produces an output. Further, a similar situation occurred at Google too but it did not shut it down. They stated that 

The network must be encoding something about the semantics of the sentence

Courtesy: Terminator

Who's right or wrong? That is where we take our AI to. The concerns regarding AI is relevant because the pace at which a machine learns is far greater than our own. The regulations must be there to prevent any extreme scenario. On the other hand, the fantasies of the dangers of AI that have become popular are at-least a decade away from realization. And in computer terms '10 years is a lifetime'.  

So we wait, cautious

What do you think. "Will our best creation be our destroyer?"