A.I. in the "Cloud"
The world seems as though it is starting to move faster and faster, and thus the demand for information and information accessibility is drastically speeding up as well. Modern computers and related technologies, however, have done a remarkable job with both creating and keeping up with the ever growing demand for data and access people need to it. Perhaps one of the interesting innovations on the scene as of late is the emergence of a new form of information sharing and storing colloquially called “cloud computing”.
The term is probably familiar to many people, but the exact logistics of what the “cloud” is and how it works probably isn’t. On the whole, the “cloud” can most simply be thought of as a network over which computing software, resources and information can be simultaneous stored and shared (a perfect example of such a network being is Internet). While such an idea may seem to demolish the personal aspect of personal computing, the silver lining of the “cloud” is it’s ability to make common resources, such as software, “apps” and data readily available to everyone connected to the network from one (or a few) centralized sources. So, instead of…let’s say…Corporation Corp. needing to buy a new copy of Microsoft Office for every computer they have in their building, they can have the software supported on one central server that everyone in the building can be grated access to use. This goes the same for personal data as well; instead of needing to store everything on your computer’s hard drive, you can just upload it to a “cloud” network and have instant access to it everywhere you get internet service! The implications of this are pretty amazing when you consider how this could transform computing on a commercial level. However, one of the most interesting aspects (in my opinion) of the dawn of “cloud” computing comes in at the personal level, specifically the personal level invested in social networking.
Think about how many people you know that have a Facebook or a Twitter account, host their own blog or a Tumblr, have a Spotify or Pandora account. That’s a lot of people, and more importantly, that’s A LOT of data that each of those individuals has generated about themselves floating about on the internet! While the “cloud” may host hundreds of apps, programs and other services, it also hosts A LOT of information about everyone one of those people you just thought about, and about yourself as well most likely (talk about really living with your head in the “clouds”)! While this may seem innocuous enough, expect for those pictures from the Christmas party on your Facebook profile (“it’s okay, I untagged them!”), media entrepreneur and commentator Adam Ostrow has another very interesting take on the “you in the cloud”:
When you think about it, a lot of what Ostrow is getting at is very true (in a rather creepy, Matrix-y sort of way). There is so much data floating around about you, and specifically you, on the internet that it’s mind-boggling. From all the conversations saved on Facebook, to the constant updates on Twitter, the photos of favorite animals and TV shows on Tumblr, catchy songs played over and over again on Pandora…any way you look at it, the internet “knows” A LOT about you, what you do and what you like. Is is so far fetched to think that a program could be designed with a algorithm in place that could allow it to take all that data about you, integrate, and “predict” what you may tweet or post or play next? Personally, I say it isn’t…and, in fact, as programs and mods that Ostrow talked about show, it’s getting very close to being possible (on a basic level of course)!
But what if we could take this beyond the basic? What if this “digital self” strewn all across the internet could be compiled, fine tuned, and modified to create an “artificial self” that, as Ostrow suggests, lives on after you die or in general just lives as you would based off what it knows about you? Think about it, a holographic you (just like Will i Am…funky!) talking and interacting with people, spurting information about what you would probably do today, what LOL cats you thought were funny and could “re-post” to everyone around you, and your latest opinion on so and so’s new album (“you’ve probably never heard of them”). While this may all seem like a simple novelty today, I think this somewhat unsettling topic taps into an even greater implication: as basis for emergent artificial intelligence.
Our usual thoughts about A. I. probably trail off the realms where cybernetic skeletons hunt for future saviors of the human race, rogue computer programs toy with humans, or sassy pixels guide us to the next waypoint on a map. All in all, A. I. is usually consider something strictly robotic and that has to adhere to what the name suggests: the emergence of autonomous thought or “consciousness” from a purely synthetic construct. Indeed, world spanning endeavors such as the BlueBrain project propose to build a fully synthetic brain as a way to best study it’s mysterious functions and, more importantly, shed some light on the elusive topic of consciousness itself. But what if this “cloud” of information could serve as the rough start of something resembling what we want a robotic organism with a metal brain to be one day? While the thoughts of a “digital persona” may be restricted in the sense that it may not be able to “think” but only predict based off previously catalogued information, who is to say that it can’t get better at this? A tweak here and there (and years of work and research later) and we could have an algorithm that takes this compiled information and extrapolates it, compares it to information and exchanges you share with friends and loved ones and employers by accessing conversations, emails and pictures you shared with them all over the internet, adapting itself and getting better and better at predicting patterns that exist within your “internet autobiography”.
Indeed many simple adaptive learning programs already exist, such as the ones worked on by the CELEST lab here at BU for mimicking and modeling speech patterns and sounds, and at MIT’s Personal Robotics Group for modeling and learning fine motor control and movements (as seen in the facial expressions that can be seen modeled by their Nexi robot). Imagine taking that iSelf and integrating it with a machine that could walk and talk like a human being and let it start “predicting” things you would say based off all the information it can access about you, and learn to mimic your words, your likes, your “thoughts” better and better. It may not be HAL 9000 or Sonny, but in a way it is a certain form of artificial intelligence.
While the prospect of a truly text book artificial intelligence is probably still more feasible in the world of science fiction than science fact, the prospect of the “digital you” is very matter of fact in the real world today. Who knows, if programs and algorithms that compile and store information about you like My Next Tweet, and ifidie evolve and integrate more and more with the web, maybe one day we could see something truly new or “intelligent” evolve out of the data stream in a way no one ever thought possible. Perhaps the concept of a viable “artificial sentience” (fans of Jane in Orson Scott Card’s Speaker for the Dead represent!) is still just as far off…and then again maybe not. If anything though, it may be a good idea to take our heads out of the “cloud” from time to time and truly consider what programs are learning about us, and what information we want to leave for our “digital selves” to compile.
How Cloud Computing Works – How Stuff Works
Role of Expressive Behavior for Robots that Learn from People – Philosophical Transactions of the Royal British Society (Biological Sciences)
A Neural Theory of Speech Acquisition and Production – The Journal of Neurolinguistics
Transforming Robotics with Biologically Inspired Models – Live Science