Prompt Images
I. Humans: A Domesticated Species
The common narrative goes that homo sapiens was a hunter/gatherer species for the first million years or so of its existence. Then came the agricultural revolution, roughly 12,000 years ago. Previously nomadic groups could finally quit the constant hustle of foraging and spearing large prey. This was all thanks to the latest clever trick that homo sapiens had figured out: domesticating plants, especially wheat.
Among the many fascinating bits of insight packed into Sapiens: A Brief History of Humankind by Yuval Noah Harari is a counterintuitive reframing of the above story of our species. Harari suggests that the common narrative is incorrect.
Humans didn’t domesticate wheat.
Wheat domesticated us.
If this sounds crazy, it’s only because we tend to think of domestication as always resulting in something more suitable to our existence.
Domesticated plants and animals for our varied culinary tastes. Domesticated cats because who has the time and resources to manage a tiger? The list goes on and on. But as Harari explains in Sapiens, there is much historical and anthropological evidence to suggest that the change from hunting/gathering to agriculture made life less suitable for humans, at least at the individual level and in the short-term.
At the level of the species as a whole, of course, it enabled us to feed people on scales never imagined, allowing for an explosion in population that would never have been sustainable on twigs and berries. But the great surplus in food for humanity as a whole actually came with a greater risk of starvation for individuals (hunter/gatherers don’t usually have to worry about failing crops). And if the greater likelihood of starvation wasn’t bad enough, agriculturalists spent much more time working their fields than hunter/gatherers ever had to spend foraging. Perhaps most surprising, Harari cites evidence that the immediate post-agricultural period was actually more violent than the pre-agricultural period (which was admittedly still pretty violent by today’s standards). [1]
Meanwhile think about things from the perspective of wheat.
Before 10,000 or so BCE, grasses like wheat were small time players in the ecological world. Then this smug ape comes along and spreads it around the world like a plague. Now wheat is suddenly—quite literally—a king maker. And those smug apes work harder, longer, and for less individual security—just to keep their wheat alive at all costs.
So yeah, maybe it’s not so crazy to think that wheat domesticated us. (By the way, these Sixth-Sense-he’s-been-dead-the-whole-time kind of moments come fast throughout the entirety of Sapiens—so if you haven’t already read it, you really should).
Why am I discussing wheat in an article purported to be about artificial intelligence? I promise I’ll get to that. But first, let’s talk about the different kinds of artificial intelligence.
II. A Tale of Two AIs
In August of 2016, Wired magazine published an interview with president Obama on the topic of artificial intelligence. Here’s Obama on the larger issues surrounding AI:
My general observation is that it has been seeping into our lives in all sorts of ways, and we just don’t notice; and part of the reason is because the way we think about AI is colored by popular culture. There’s a distinction, which is probably familiar to a lot of your readers, between generalized AI and specialized AI. In science fiction, what you hear about is generalized AI, right? Computers start getting smarter than we are and eventually conclude that we’re not all that useful, and then either they’re drugging us to keep us fat and happy or we’re in the Matrix. My impression, based on talking to my top science advisers, is that we’re still a reasonably long way away from that. It’s worth thinking about because it stretches our imaginations and gets us thinking about the issues of choice and free will that actually do have some significant applications for specialized AI, which is about using algorithms and computers to figure out increasingly complex tasks.
As with all of Obama’s verbal answers, it’s long-winded, it’s as polished as a piece of prose that’s gone through several back-and-forths with an editor, and it’s spot on. [2]
One of the biggest issues we have in thinking about AI, as a society, is our tendency to lump different kinds of AI together. What Obama refers to as generalized AI and specialized AI might be translated, respectively, as “AI that thinks like a person,” and “AI that can do a particular task really well.” Specialized AI can already play chess, Go, and Jeopardy better than humans. Soon, it’ll probably drive cars better too. But there’s no generalized AI that can play those games AND talk politics AND decide who played James Bond better: Sean Connery or Roger Moore.
Compared to generalized AI, specialized AI is pretty dumb. But then again, so is wheat. And that didn’t stop it from domesticating a species that was vastly more intelligent.
III. Chief Operating System
Towards the end of his second book, Homo Deus, Harari describes one possible future that may await mankind. This future is dominated not by sentient AI who come to think of us an existential threat, but by a host of simpler algorithms that don’t really “think” much at all, except in a very narrowly defined sense.
Harari considers the case of Deep Knowledge Ventures, a Hong Kong venture capital company that appointed an AI to a position on its board. That is, the AI (named Vital) votes on what companies to invest in alongside its human board members. Is it so hard to imagine, Harari asks the reader, that one day specialized AIs may supplant all the board members of a company? That the top position in any company might eventually be the COS (Chief Operating System)?
If we extrapolate further, is it really out of the realm of possibility that specialized AIs might one day own companies?
Harari is quick to remind us that once upon a time people invented the idea of corporations, giving legal protections to an otherwise “artificial” entity. Who is to say we might not do the same thing again, giving algorithms, even relatively dumb ones, legal ownership of corporations? If they can produce better results and do so with less corruption, we might find it hard to justify not putting them in charge of our economic affairs.
Consider how much control we’ve already given over to companies like Facebook.
We trust their algorithms to curate what news we see, which memes we’ll laugh at, and whose lives we’ll covet. And the company claims to know us better than we know ourselves, based on analyzing our history of ‘likes.’ Again, this claim doesn’t require Facebook to have the most advanced artificial intelligence one could imagine. In fact, I bet I know some of my friends better than they know themselves (and vice versa) after scrolling through a sampling ‘likes’ on their feeds. Even a dumb algorithm with unprecedented access to information about 2 billion of the world’s citizens is a force to be reckoned with.
Surely you can feel it happening. 12,000 years after we let wild grasses fence us in to a narrow physical space, we have let the algorithms define virtual boundaries for our lives.
IV. Why?
All this begs the question, why have we done this to ourselves?
We could ask a similar question of our pre-agricultural brethren. Why trade in the culinary perks and leisurely pace of the hunter/gatherer lifestyle for the monotonous diet of a single grain + tubers, and 12-hour days of back-breaking labor? I’m not a historian or an anthropologist, but I suspect that it began as people experimenting with growing their own food, and succumbing to the same fantasy of every Park Slope hipster with a window box of basil and thyme. Life would just be easier if I could grow all my own food.
It was only later that all the downsides of the agrarian lifestyle—famine, disease and an 80-hour work week—became apparent. And by that time they were in too deep (literally, into the ground) to back out. [3] In the meantime, the agricultural revolution allowed humans to feed population levels that could never have been sustained previously, thus giving our species yet another advantage over the rest of the biological world.
Turning back to our new algorithmic manifest-destiny, it isn’t hard to see why we’ve let the little buggers infest our lives. We started by experimenting with apps and Echos, and in the process we’ve discovered a new kind of fantasy: Life would be easier if Alexa did everything for me.
It’s only now, a decade or so into this experiment, that many people are looking up from their iPhones to ask about the downsides. Shortened attention spans. The sense of drowning in too much information. Political propaganda spread like influenza.
But it may be too late.
And even if it’s not too late to walk some of this back—to sign off and unplug. It may be inevitable in the long-run anyways.
Because if we zoom out from the level of our particular species, we see that life itself is in a scorched earth war with entropy. Complexity emerges as life figures out ever more ingenious ways to use energy to keep entropy at bay.
Until now, humans have been the most complex systems ever devised, creating order among the chaos at unprecedented levels. But there’s no good reason to suspect that our minds can indefinitely adapt. Life might decide that in lieu of machines with superhuman intelligence, humans + dumb algorithms will do just fine.
Which of these artificial intelligences is the greater threat to humanity?
Look, I’m not saying that this is our destiny. Maybe we’ll beat back the algorithmic invasion somehow, or find ways to thrive as a species in spite of it. I am just pointing out that if AI does take over humanity, our overlords might be more Microsoft Paperclip, than T-1000.
Footnotes:
[1] It should go without saying that I’m neither a historian nor an anthropologist. Harari is a historian by training and leans heavily into the “hunter/gatherers were happier than agriculturalists” idea. But I’d be misleading you to suggest his view is sacrosanct among experts. If you’d like an idea of the complexity of trying to answer which is the “better:” a hunting/gathering lifestyle or an agrarian lifestyle, I suggest you read this great (and nuanced) answer on Quora. Regardless of the ultimate answer to that question, agriculturalism had its downsides, which I’d guess most people don’t tend to think about.
[2] One can’t help but imagine what Trump might say if asked to comment on the state of Artificial Intelligence. The computers are great. They are really, really good. I know a lot of them and I think they are really great computers. I do worry, that, maybe there are some bad ones. But mostly I think they are doing a great job and they get a bad rap from the media.
[3] To be fair, many thousands of years later this means of obtaining food has made life easier for a majority of the world’s citizens. But it’s taken some time, and in the meantime we had to go through drought, plagues of locusts, and potato famines. As Harari points out, there is zero chance that our ancestors would have chosen the agricultural path because of its benefits to their great-great-great-great-great-great-great-great-great-great grandchildren.
RELATED
