Almost as exciting and frightening as AI itself is comes the speed with which Google is pushing it’s development. Just last week I discussed in my blog post, “Let Google Answer Your Emails” the gains Google is looking to make in understanding language and intent and how they’re pulling us into their learning equation. With today’s announcement we’re taking this one step further.
The technology is called TensorFlow and they’re making it open source as part of the Apache 2.0 license. TensorFlow is used by Google to train its systems to understand more-and-more about how to understand the world around it. From recognizing images to translation, its only function is to learn how to “think”.
By making TensorFlow open source Google is giving away access to a massive technology. But let’s quickly look at that sentence one more time and let’s bold the most important word:
By making TensorFlow open source Google is giving away access to a massive technology.
What’s critical to understand in this is that Google is not giving away the technology to run in isolation but rather allowing others to use it for different purposes and expand on it with the data being fed back to Google. When I discussed the email answering issue on my radio show I likened Google’s AI to a 4-year old. With the addition of the email system they were looking to let others give basic conditioning rules in a controlled environment. Basically … they were taking their little algorithm to church to hear about the ten commandments. With TensorFlow they’re upping their game and sending it off to school – a far less controlled environment where different people will say and teach different things. The goal here isn’t to teach the algorithm how to behave but rather how we behave in the wild.
One can expect to see TensorFlow deployed across a wide array of sites and applications from kids video games and photo recognition to dating and porn. Essentially, Google will get what they want and that’s a picture of the fuller scope of how humans interact in different environments and they’re letting us steer the way.
To be sure, this is not a benevolent act. Google needs this data to push their AI further. There is only so much they can learn through crawling the web and emails and programmed by their engineers. To further AI they needed to push the boundaries outside of what they themselves can collect and into the world at large. Today they have done that and the data they will collect will be massive. The question as to whether this is good or bad will depend on who you ask but there’s no doubt that the push for true AI is on and I’m putting it in the “less than five years” category with its current trajectory. But as that happens let’s remember the words of Stephen Hawking:
“A superintelligent AI will be extremely good at accomplishing its goals, and if those goals aren’t aligned with ours, we’re in trouble.”
“You’re probably not an evil ant-hater who steps on ants out of malice, but if you’re in charge of a hydroelectric green energy project and there’s an anthill in the region to be flooded, too bad for the ants,” Hawking added. “Let’s not place humanity in the position of those ants.”