henry kissinger’s misplaced warnings about artificial intelligence

Kissinger’s concerns about machines displacing human curiosity and turning our brains into useless, gray jelly unsuccessfully mine the same territory as many other technophobes with roughly the same results.

robotic geisha ghost in the shell
Robotic geisha from Ghost In The Shell (2017)

Henry Kissinger believes that artificial intelligence will outthink humans into irrelevance according to his lengthy piece in The Atlantic, echoing a common theme of many technophobes who can’t help themselves from quaking at what amounts to comic book science and engineering. And in no small part because he’s a famous public intellectual regardless of his, shall we say, mixed reputation in the foreign policy community, he’s been given a massive platform to spread his fears to those who might not know much about computers and AI, guaranteeing that his vision of a slow-motion apocalypse of human thought will be quoted for years.

Now, while one should always at least try to be polite to one’s elders, we also shouldn’t hold back from blunt criticism. With that in mind, let’s point out the first problem in Kissinger’s essay. It’s basically an equivalent of a padded term paper about AI from someone who dove into a few basic Wikipedia articles about it and with the good sense not to start it with “The Oxford English Dictionary defines intelligence as” which, let’s face it, shouldn’t be the opening line in any piece of written content. (No, invoking Webster’s is not a valid loophole, and shame on you if you thought about it.) This self-admitted lack of subject matter knowledge already renders the essay and its conclusions highly suspect.

The next problem comes when Kissinger follows the accomplishments of artificial neural networks down a path every technophobe and philosopher along with their second cousin twice removed has stumbled. Since computers are getting better at us in certain tasks, he reasons true to the cliche, one day they’ll be able to do things we couldn’t even think of and eclipse us in all aspects of science and technology. He then abruptly pivots to the other trope on loan from the Luddite museum, the idea that techies and technophiles are allowing technology to do all their thinking for them and losing the ability to critically evaluate claims. These two arguments make up the heart and soul of his thesis, and both are completely off base.

First, let’s tackle the power of artificial neural networks, or ANNs. It’s true that we’ve built ANNs to do a lot of amazing things, and do them better and faster than humans. What isn’t true is that they constitute a step in artificial general intelligence, or AGI, the notion that we can build an entire artificially conscious and intelligent entity like Ultron in the Avengers series. Kissinger is basically arguing that because statistical algorithms don’t have limitations on short term and long term memory and were built to do math much faster than us, comic book science must be on the horizon. It’s the equivalent of thinking that because your calculator can multiply 8,976 by 3,472 in milliseconds, it will one day supplant you.

For all their prowess, ANNs are simply fancy calculators based on statistical formulas we devised and are using to crunch through tons of data for predictable, well-defined tasks so we don’t have to do it ourselves. They’re just tools, impressive tools, but tools nonetheless, and comparing them to a hypothetical AGI is like comparing rocket engines to warp drives. We have rockets and we know how to build and improve them. Meanwhile, we don’t know if a warp drive is possible, our ideas for how to make one work are just starting to take shape, it might have some very destructive side-effects, and if we get one working, odds are, it will be nothing like what we imagined it would be with our currently limited understanding. And incidentally, ANNs would very likely play a role in building warp drives by accelerating how quickly we can process data from complex physics experiments in particle colliders and outer space.

All this brings us to Kissigner’s second argument, that relying on computers to crunch data will turn our brain and critical thinking skills into mush. At its root, this is little more than Luddite condescension, telling the computer literate that it doesn’t matter how good they are at finding information, they’ll never be deep thinkers capable of delving into complex issues because they apparently have no capability to put data into any sort of context. Or as Kissinger puts it…

Users of the internet emphasize retrieving and manipulating information over contextualizing or conceptualizing its meaning. They rarely interrogate history or philosophy; as a rule, they demand information relevant to their immediate practical needs. In the process, search-engine algorithms acquire the capacity to predict the preferences of individual clients, enabling the algorithms to personalize results and make them available to other parties for political or commercial purposes. Truth becomes relative. Information threatens to overwhelm wisdom.

From an academic standpoint, one could argue that techies are being dehumanized here, their ability to do more than regurgitate what is said on the screen stripped from them. At the same time we do have to acknowledge that the part about truth becoming relative and information overwhelming wisdom are very valid concerns, though not for the reasons Kissinger imagines. In his attempt to indict computer users as lacking the faculties for critical thought, he bumped into confirmation bias, a well known flaw with human psychology in general, and the complete lack of forethought by social media platforms on how to handle the information with which they are routinely deluged.

You see, with every possible viewpoint on an issue competing for people’s attention, many will simply choose what they want to be true, just as they have since time immemorial. With few gatekeepers and no screening for truth or quality, social media gives them a handy excuse to engage in confirmation bias all day long. But this is hardly a new development when we look at the history of tabloids, yellow journalism, and the fact that fake news, urban legends, and political propaganda have all been with us since the start of recorded history. What we’re seeing today is nothing new, it’s just happening a lot faster because the world is now mostly wired and content travels quickly across the web.

That’s why it’s dangerous to assume that social media or the internet exposes us to people who are unwilling to exercise critical thought or have the manners of feral troglodytes. The problem isn’t the computer they’re using, it’s their impulse control. Hundreds of millions of people manage to use the web, computers, the products of ANNs, and robots to live productive lives, stay sane and informed, and harness the information they learn thanks to all this technology to build everything from new companies to volunteer movements to tackle the world’s ills. We could argue that we need to teach people more about how to have a healthy relationship with technology and capitalize on their innate curiosity more often, but let’s not pretend that computers will somehow undermine us in this effort. On the contrary, they’ll help us.

Kissinger would find it instructive to consider the response of doctors to an artificial intelligence created by IBM and designed to help them treat and cure cancer. Far from taking the machine’s outputs at face value, they rigorously tested it and pointed out the many flaws in its real world implementation. It’s the same story in every field we’re using AI to speed up, automate, and scale up work. The minute you dive past the hyperbolic Luddite rhetoric, it becomes obvious that humans are very much in control of their digital creations.

There are plenty of legitimate concerns about how we apply ANNs and social media, but technology turning us into passive meat with eyes and wreaking havoc on the world with its unguided whims is absolutely not one of them. But I suppose it’s convenient to look at the all the terrible things humans have been doing to each other with technology over the past few years and blame an inhuman culprit who won’t object and defend itself, excusing our own worst impulses and flaws with abstract meanderings through a medley of pop sci news bites. And fundamentally, that’s exactly what Kissinger and his fellow technophobes are doing.

# tech // artificial intelligence / social media / technophobia


  Show Comments