What everyone assumes Google's plans to be
For those of you who weren't already aware, Google - which just finished purchasing a whole host of robotics firms - this week bought AI startup DeepMind for the tidy sum of $500 million. You can be forgiven for being a little nervous about the news. After all, Google did just finish buying a whole laundry list of robotics firms; as such, a great many media outlets are of the mind that it's looking to build brains for its new bots, and that, in the process, we might see the birth of a real-world Skynet (or at the very least, mass unemployment).
It's certainly possible. I wouldn't discount the possibility of Google trying to be the first company up to bat where the advancement of the robot world is concerned. After all, they're not exactly known for being sluggish in the field of innovation; their not-so-secret secret labs at Google X always seem to have one or two science fiction-esque projects on the go. A real-life, fully autonomous Android seems like just the sort of thing Google would want to work on.
As cool as such a technology sounds, however, it's also cause for some concern. As you may recall, robots are inevitably going to better our lives as the technology advances, but there's a good chance they'll cause a great deal of suffering before they do so. I'm not talking about some Terminator-esque machine uprising, either: what I'm actually referring to here is mass unemployment, economic instability, and a complete upheaval in the way capitalism functions.
None of those thoughts are very comforting, are they?
Of course, there's another possibility regarding Google's plans for Deepmind, as well. In spite of all the innovative technologies it's pumped out over the years, Google is, at its heart, a search engine organization. And as it's no small secret that it's constantly looking to improve its search algorithms and make them smarter, better, faster, and stronger; well...there's a good chance Google's plans for DeepMind involve not robots or androids, but search engines.
This thing's terrifying enough without being able to think, too.
Let's take a look at the evidence, otherwise this is just meaningless conjecture. At the time of acquisition, Deepmind was working on three projects: "A game with very advanced game AI, a smarter recommendation system for online commerce and something to do with images." Two of those would work very well in the context of a search engine, and the third probably doesn't have much of an application in robotics intelligence.
Further, DeepMind is seen by Google more as a talent acquisition than a technology acquisition. Although it's gaining access to a few of DeepMind's patents in the process, the search giant is more interested in what the team can offer the organization than it is any technological breakthroughs accomplished by them. That doesn't exactly speak strongly in favor of a robot apocalypse, does it?
The strongest evidence that Google's not planning to build artificial brains anytime soon, though? The DeepMind team isn't going to be working with Google Robotics head Andy Rubin. Instead, they're going to be reporting directly to Jeff Den, Google's resident search guru.
Alright. So there's a good chance Google's not trying to take over the world with the DeepMind purchase. But what are they trying to do?
Good news: Google isn't trying to conquer the world.
Recode reports that Google's purchase of DeepMind is very likely tied to the analysis of unstructured - or "big" - data. Advances in machine learning have already allowed for the development of algorithms capable of sorting through the terabytes of chaotic information, but the process isn't quite perfect. DeepMind's own advances in Machine Learning could very well help Google to further refine its algorithms.
When Google purchased DeepMind, people all across the net expressed a combination of fear and excitement. Some predicted Google might be putting together an ethics board for AI. Some even went so far as to posit that Google was building the machine intelligence equivalence of The Manhattan Project; that they could either work with the military to create a literal robot army or bring about mass disruption in the job market.
None of these are terribly likely, at least in the short term. It's fairly clear - at least to me - that Google's not looking to build a terminator; instead, it's trying to make its already-formidable search technology even more powerful. That said, it's certainly possible that Google might also be working on AI on the side.
"Google uses machine learning in every nook and cranny of what they do," explained University of Washington Computer Scientist Pedro Domingos to Recode "Larry Page and Sergey Brin don't say it, but they want to solve the AI problem. They really do want AI to come true."