Whether we like it or not, Algorithmic Intelligence is here to stay. Just like with genetic modification (of both food and humans), you couldn’t stop it even if you wanted to. And to be fair, it genuinely has the potential to make life better for everybody. [EDIT: see this Atlantic article for a better summary of these issues.]
The challenge is that there’s not going to be a straight line from here to there, and history suggests things will probably get worse — potentially, A LOT worse — before the benefits outweigh the costs.
To be clear, I’m not a pessimist by disposition. I believe problems in life are both inevitable and solvable, which means we shouldn’t fret unnecessarily when a new one appears. But we do need to understand the typical course:
- A new invention or discovery
- Optimists (typically liberals) foresee what we CAN do with it and assume therefore we will
- We do that a little, but we also do lots more bad shit than anything else — often completely unanticipated shit. For example, the internet facilitates commentary like this, but measured by weight it’s mostly porn. By far. After that, it’s gambling, cat pictures, selfies, scams, misinformation, and organized crime, both private and public.
- Things get better for elites but typically a little worse overall and the novelty is soberly re-examined for what it really is versus what we wanted it to be. We’re at this stage now with social media/big data.
- Bans and prohibitions are enacted, addressing symptom rather than cause
- The wealthy quickly claim beneficial exceptions, often in isolated “walled gardens” separate from the rest of us
- Generations of inequality and/or outright suffering follow, such as after the invention of the steam engine, which made products more affordable for everyone but also ripped limbs from children for decades
- Some time later, people start to contemplate the responsible governance they should’ve considered at the point of irrational exuberance
- Changes are introduced gradually over the constant objection of political conservatives, and slowly, mostly by trial and error, life gradually improves for everyone but the poor, who are pretty much ignored by Right and Left alike.
For me, the problem with AI isn’t just that it introduces another social asymmetry, such as what we’ve seen repeatedly in history going all the way back to the invention of agriculture, but that it introduces the ultimate asymmetry.
In the past, the wealthy and powerful were always — at some point, once you got far enough down — dependent on the poor and working classes: to grow food, to staff factories, to fill armies, to clean house, etc. That meant there was a floor to human suffering. Things could only get so bad before workers went on strike or the population rebelled and so introduced a “correction” in economic terms. This is the fodder of history, the dates of which you were forced to memorize in school.
[For those who want a really great overview of those kinds of forces across the span of human history, read William McNeill’s Plagues and Peoples.]
AI — particularly a conscious Artificial Intelligence, which is different than the kinds of Algorithmic Intelligence that get the label these days — has the genuine potential to render the great mass of people not just superfluous but an outright burden. There will be no reason, in a politico-economic sense, for most of them to exist since neither their labor nor their vote will empower the ruling class.
When AI produces the food, manufactures the goods, and comprises the military, there will no longer be anything to stop the powers-that-be from reproducing and extending the “solutions” that have repeatedly occurred throughout human history: from Stalinist Russia to the pogroms of medieval Europe to the Khmer Rouge to the Rwandan genocide and on and on, right up to what’s going on in Myanmar at the time I’m writing this.
Wherever we go, there we are. With whatever tools we invent, the hand that holds them is still a human one.
Futurists in the 1930s, for example, predicted that one day, with the invention of a robotic labor force, men would no longer need to work. Their erroneous assumption was that as a result of robotic production, resources would be shared by all rather collected at the top. There is no end to human greed.
AI might eventually reach such ubiquity that all (or most) humans alive will benefit. That’s definitely possible. But if so, there’s still the question of the path it takes to get there.
Imagine traveling back in time to the year 1900 and telling a group of people on the street:
“Okay, look… I gotta be honest. Things are gonna be bad. Really bad. There’s gonna be a horrible global war in a few years that will kill unprecedented numbers of people, not just through violence but also famine and disease. After that, a horrible influenza epidemic followed in a decade by a massive economic depression will put huge chunks of you out of work and force you to move from your ancestral homes. Many will die, and it will only end because of another world war, this one even larger and more devastating than the first (if you can believe it), and which will see the invention and use of weapons capable of wiping out the human race. In the aftermath, the European colonial empires will retract, which sounds great, but they’ll leave a power vacuum, and the developing world, from South America to Africa to Asia, will experience repeated waves of civil war, ethnic cleansing, and famine . BUT… there’s a silver lining! For those who survive, things will actually get better than they are now. Democracy will spread. Basic social safety nets will be introduced, including a minimum wage. Women will get the vote and racial integration will be the law of the land, if not always the practice. Health care will improve as well as opportunities for education and home ownership. So chin up! Smile on your faces! And back to work!”
Those who made it through all that — you and me—are the beneficiaries. It’s all just some bad stuff that happened before we were born, like the Inquisition.
It’s great to say “We’re gonna have universal basic income,” except we’re probably not — not anytime soon anyway. We couldn’t even get universal health care coverage in this country under a nominally Democratic administration. The tax bill under consideration in Congress at this very moment tilts the complete opposite way.
Maybe, after we go through all the pain, people will agitate for something like UBI get it passed. If so, that will be great for the people alive then, and they’ll look back on the changes that are unfolding around us now as something bad that happened before they were born.
But as a practical solution, high hopes don’t really address what’s coming.