We looooove the idea of having an Artificial General Intelligence (AGI), don’t we? And even more so, we love the idea of what a human-level AGI could potentially create: a Superhuman Intelligence. I sure do love that idea. It’s like making a God who’ll be more or less tangible, giving us world-changing results, fix all our problems, and blow up balloons the shape of molecular structures at kids’ birthday parties (and yeah, possible destroy us all but whatevs). Happy happy. My problem is with WHO most likely will create and operate the AGI.
I’m as excited as anyone with a Utopia AI or Beneficial AI mindset (see Max Tegmark’s description of that in his amazing book Life 3.0). While there is zero consensus as to WHEN we’ll have an AGI – or several? – many AI researchers who are deep in the trenches do believe it’ll happen at some point. Optimistic guesses from folks like Ray Kurzweil and others put the rise of the first AGI somewhere around 2030 or even sooner, while more pessimistic folks put the date somewhere closer toward 2070 or later. These numbers are meaningless because nobody really knows. We don’t know if Moore’s Law holds, we don’t know how much Network Effects really play a role in the development process, we don’t know if we’ll be able to create that AGI with our “traditional” computing or if it requires quantum computing, etc. Of course some super smart folks claim they do know. They don’t. And educated guess is still just a guess. Nobody has all the variables or can foresee technological developments with certainty. They all guess, they wish, the assume. But they don’t really know. And that’s all fine.
I’m in the camp that believes we’ll be worshiping our very own
God AGI sooner rather than later. My gut feeling and frankly, my wish is for it to be active around 2030-ish. Of course, I don’t know either. But the real point is not when it’ll happen. The problem I have with all this is WHO will have it.
While we all dream of having an AGI that is built by a group of super smart, super benevolent scientists who get there because of sheer determination and programming smarts, if we view this issue by stripping away hopeful thinking then we must face the most likely possibility: the AGI will be created by one of 3 companies, all of which are gigantic for-profit entities and 2 of which are notorious for grabbing as much data as possible on us mortals and monetizing it.
Who will create the first AGI
You already know who that is. The first AGI will most likely be owned by Google or Facebook. Distant third contender is IBM. Alternatively: China. doesn’t matter which Chinese company or University or “research group” because they’re all in the Chinese governments’ pocket.
You already know this, deep in your gut. It’s time we all admit this to ourselves. Our beloved AGI, our digital lord and savior, will most likely be built and owned by a private, data-mining mega company. It will run our lives. It will determine how we live, work, love, and die. We will have no oversight, no control, and no way to protect ourselves from it. We’ll have no say over how we best put that AGI to work for us.
To be clear – ALL of these companies already use artificial intelligence, and have been for a long time. Machine learning and Google and Facebook are inseparable. But that’s narrow AI. I’m not saying that whoever has the most narrow AIs running things is the one that’ll ultimately “graduate” to Artificial GENERAL Intelligence land. Having the most narrow AIs is not necessarily the most important factor to win the AGI race.
A Google AGI will probably be really good at knowing where every person on earth is and what they spend their money on, then sell them more of it. They’ll dangle the occasional “beneficial/ do no evil”-type AI solution to make us all feel good about it and keep us docile. Google’s many, many proverbial fingers in multiple disciplines, including science, autonomous vehicles, and many others, further gives them the edge here. When time allows I’ll elaborate on why Google is likely the one to achieve AGI supremacy before Facebook.
Similarly, Facebook’s AGI would be really good at putting humans into silos, assigning a social status value to everyone, and selling that information to anyone with a sack of shekels. I see Facebook as being less likely than Google to come up with the first true AGI, considering the more narrow range of data FB collects (as far as we know). It’s also less trustworthy (see recent slew of scandals) and offers fewer hooks into the business and enterprise world compared to Google.
When Facebook blatantly sold confidential user data (Cambridge Analytica is just the tip of the iceberg), there was practically no real punishment for the company. Zuckerberg’s spy agency may get a slap on the wrist, he may squirm in his tiny seat at the Senate hearing, but in the end nothing is going to happen. We can expect more of the same if Facebook is the one that wins the AGI race.
My personal dislike and bias against Facebook is tangible here but so far nobody has been able to prove to me that FB is not the devil Google at least does give truly useful tools to mankind. (Full disclosure: I used to be a Google fanatic. Remember Gmail invites? That’s when my fanaticism for Google started. I bribed my way into Googel Voice and used my GV number religiously for everything. Google Docs? OMG, what a life saver. My love for Big G didn’t end until the Edward Snowden disclosures. I was a sheep. Probably still am, just no longer one that follows Google blindly).
Why only Google or Facebook?
An AGI NEEDS as much data – different types of data- as possible. It also requires lots of money, lots of very smart people, and as little public or governmental oversight as possible.
And the 2 companies with that kind of data, money, people, and secrecy are- Google and Facebook. Both can do what they want, when they want, and there’s nobody who can truly put a stop to it, or even find out exactly what kind of data they have, or on whom, or what they do with it.
As far as we know, nobody has more data than Google or Facebook, with the possible exception of the Chinese government. Amazon may be a very distant 4th contender, not the least because of its recent remarkable push to put a listening device into every home and car (Alexa). But the type of data Amazon collects is too narrow, and despite the occasional dabbling into non-sales fields (facial recognition software for law enforcement. Not kidding, look it up), Amazon doesn’t seem to be positioned as well as Google or Facebook to create a true AGI.
So if Google or Facebook run our lives, who can step in if – no, when, they overstep the line and start hurting us? Our governments are useless. Again, see the “slap on the wrist” comment above. This is what happens when a company is too big and powerful to really punish if (when) they do something that’s against the law, or against the better interest of the general public. Google and Facebook have enough lawyers and money to squash any serious punitive measures. They also have enough positive clout with people – hey, I get free email and docs and can post pics of my dinner online! – to get away with pretty much anything. ANYTHING.
IBM has been in the AI space for a long time, and has massive amounts of data as well. However, IBM focuses more on the enterprise sector. They don’t provide and therefore don’t snoop on billions of email addresses, spreadsheets, maps and consumers’ movement data. Google does. Facebook is the king of knowing humans and their relationships and likes and dislikes, even better than an individual does. So Google and Facebook, having the most relevant data to eventually build an AGI, and having more money than they know what to do with, as well as the smartest developers on the planet, are best positioned to create the first AGI.
But I believe IBM is a distant third. They do have incredible people, more patents than anyone in the world, and lots and lots of data. But it’s mostly enterprise-related, medical, and other non-personal data. Did you know that IBM actually had a quantum computer before Google did? It’s true. Google’s new quantum computer is “better” , but that may not be the case much longer. IBM has some of the most innovative hardware and software developments in the world. Watson is seemingly everywhere, and since its win at Jeopardy in 2011 (sawy, Ken J) it’s become a household name and often what people picture in their minds when someone says “Artificial Intelligence”. I wouldn’t be surprised if IBM announces a quantum computer that beats Google’s in the next 3 to 5 years. If they don’t come up with a better quantum computer, they most likely will announce some other major milestone with Watson because, well, Watson has been kicking ass and getting better and better at everything it’s trained on every year. IBM was smart enough to realize after Watson’s Jeopardy win that the Watson platform and brand could be much more than a novelty AI. By now, Watson is an AI brand (really, a collection of many, many AI systems) in the fields of medicine, legal, enterprise management, and more.
Too long, too unscientific, too much gut feeling – what are you saying?
Let’s summarize my ramblings: the two most likely companies to come up with an AGI first are Google and Facebook. Big G probably has the edge, considering the vastly different data points they have and continue to gather. Facebook is second, mostly because the type of data it collects. But then again, we don’t fully know exactly what data FB actually collects and works on. We know what they want us to know.
I’d personally prefer IBM to have an AGI over Google or Facebook because, if history is any indicator, Big Blue is less likely to use it to somehow fuck us all. I have no evidence and I’m only speculating based on my gut, but I feel like IBM’s AGI will likely be one that comes up with new medicines, climate change predictions and solutions, and similar helpful stuff.
So what about China? It’s very possible China will actually beat us all in the AGI race. I’m not going into what kind of a disaster this could prove to be. I’m both in great awe and deeply scared of China. You should be too. Whether Tencent or Alibaba or any of the other Chinese tech giants or Universities get their AGI up and running first doesn’t matter, because it’ll ultimately be controlled by the Chinese government. Would that be even worse than having Google or Facebook run the world with their AGI? Possibly.
Neither scenario makes me feel comfy cozy though. And that’s why I believe we, as a society, need to be very careful about whom we allow to develop AGIs and beyond. But do we even have the power to control this?
Believe it or not, this is a very short, non-scientific observation that I intend to elaborate on farther down the line. I have some very specific thoughts on Google AGI vs Facebook AGI. If they let me live that long, I might just share them in detail.
Why are you saying “the AGI is near”?
Click bait. I have no idea if it’s near. Nobody does.