If you haven’t seen Star Wars’ Revenge of the Sith, let me recap: The above scene occurred at the transformation of Anakin Skywalker into Darth Vader after he set up the killing of Jedi Master Mace Windu (Samuel Jackson). Here, he asks, “what have I done?”
That question brings me to recent events that make me wonder if artificial intelligence researchers are beginning to (or ever) ask themselves that same question.
Here’s what happened: Elon Musk, Bill Gates, Stephen Hawkings and others have been very vocal about the risks posed by artificial intelligence. Recently, Musk and Facebook’s Mark Zuckerberg have been spatting about these risks. To Musk, Zuckerberg doesn’t know a lot about AI and according to Zuckerberg, Musk is being irresponsible by suggesting that there should be more regulations on AI. But then something happened that should raise everyone’s eyebrows.
Recently, Facebook had to shut down one of its AI systems because it – get this – started communicating in a language that it created from scratch, that humans didn’t create or even understand. While it started out using English, it later created a system that only AI systems could understand.
Researchers at Facebook shut down the system and forced it to use English. Hmmm…
Ok folks, not sure about you, but I find this really creepy. And yes, I know, there was no death and destruction that resulted from this language, and no, this wasn’t the first time this happened. But I’m still queasy when machines create their own language that we don’t understand.
So, for this round, I’m gonna give the win to Musk and Co. and hope there will be appropriate safeguards in place for these increasingly advanced and powerful technologies.