The Opposite of A.I.
Lay off the dystopia. Technology does not have to destroy in order to create.
It’s said that if you have a talking horse, nobody cares to listen to exactly what it says. It’s enough that the horse talks at all. Artificial intelligence, or the hype all around it all, is a bit like that talking horse.
Every screen is full of stories ripe of promises for what A.I. will do for us. The stories too portend the threat of lost jobs, disrupted companies… visions of entire industries soon to be destroyed.
Are you all hearing that, too? Or is it the sound of another talking horse?
It’s been a decade since IBM’s Watson promised a revolution in cancer treatment. Ten years, too, since Tesla promised an A.I. “fully autonomous” self-driving car. The real progress is, of course, just around the corner — a month or two, a year at most away.
“Just wait, who knows,” they say. “So much can happen in a year!”
Ain’t nobody got time for that. Let’s ask this horse some questions, see what it has to say.
What Does the Horse Say?
I’ve marked few improvements in the workaday places where I’d expect more impact from A.I. My bank’s fraud detection, for example, is blissfully unaware of even the most obvious patterns in my purchasing behavior.
Spell and grammar checking on my phone is famously ducked in the most obvious ways imaginable. Numbered lists in Microsoft Office…. well, everything in Office has gotten steadily worse in the last 20 years, as far as I can tell.
One of the drawbacks to having a talking horse, I think, is once it starts talking one probably grows reluctant to ask it to do any unpleasant work.
I’ve seen news stories that fashion models will lose their jobs to A.I. — I very much did not see that coming. I’ve seen wholesale loss of jobs predicted for writers, editors, truck drivers, marketers, accountants, pilots and lawyers (these from 2016!), and even doctors, teachers, and nannies.
Honestly, I was doing ok until this part about the nannies:
The question is how good can A.I. become, really? A lot of the nannies my neighbors have are really good at taking care of the kids, but they don't necessarily have the skills to teach the kids a computer language.
So, nannies are safe as long as they can code. Got it.
Don’t be Coarse, of Course
Here’s a fun tip. When you’re reading these stories, try replacing the term “A.I.” with the words “powerful yet unpredictable computer software.”
Like this:
Millions of beautiful models will lose their jobs, made obsolete by digital versions generated by powerful yet unpredictable computer software.
Among the jobs most likely to be replaced by powerful yet unpredictable software are teachers, school psychologists, and judges
Oh, now we’re having fun! This always makes me feel better.
“Bye bye models”, “Adios animators!” I this kind of drivel every day from someone who has just tried some new AI tool from Silicon Valley and I feel angry, sad, and embarrassed for them.
It’s quite coarse to celebrate the prediction of the loss of thousands of jobs. And it’s very stupid and ignorant to do it on the internet so that, in the unlikely event you are correct, the jobless masses will know exactly who to blame.
Wisdom Bottled Right from the Source
Not everything from the Internet has been positive. There’s “short form video”, porn addiction, social comparison, and the vexing slang terms “based” and “bae.”
Computers, technology, and the internet have made work safer, enabled us to live longer, created jobs, created wealth, and improved access to democracy around the world.
Technology has often, but not always, led us to productivity increases even as the quality of our work improves. This should be the focus of our optimism1 for A.I. — not the gleeful promise that it will drive the exploitative excesses of capitalism to ever-greater heights?
If your own business was “disrupted” by a giant internet company like Amazon or Uber, if you worked in a book store or a travel agency, you might see all a bit differently. I have no doubt that A.I. will change the world significantly, as much and as quickly as the internet, PCs, and smartphones did — maybe more, maybe even faster.
But are we really on the verge of eliminating millions of jobs overnight?
Points To Reinforce
Let’s travel back in time to 2005 and take a look at the impact of sudden technological disruption. Smartphones arrived and disrupted taxis. Things look very different in 2005 than they do today. There’s a narrative that the smartphone revolution destroyed millions of jobs. Is it true?
Strangely, in a sense there are actually far more taxi drivers now than then. There are actually now 10X the number of professional drivers. Of course A.I. promises (and promises and promises and promises, mostly to its investors) to replace them too, but for now we are left to argue about whether technology destroyed the taxi industry or created it.
Predicting massive job loss from A.I. is like standing in 2005, looking at a smartphone, and saying “bye bye taxis.” Even if you’re right, you’re wrong. Also irrelevant, and annoying. The predictions themselves are a fool’s errand.
What we gain or lose from technology must be understood on a human scale. No matter how powerful or unpredictable the mysterious tech wizard seems, its greatest ability is to empower human talent and ingenuity, not replace it.
This goes for the harms as well. Measuring the impact of social media on print media, for example, is useless without also measuring the damage it’s done to the concept of the truth or the impact it’s had on our mental health.
The transistor, personal computer, and digital photography all eliminated earlier fields and replaced them with vastly larger industries. Their impact on our lives, our behavior, and the trade in our marketplaces was much greater than the impact measured on individual people, professions, or careers.
To those of you who are taking so much pleasure writing and commenting on the impact of A.I. on the workforce of the future, I ask for some grace. My friends, there is something I’d like you to read, and it’s called “the room.” Is now the right moment for you to gleefully predict the loss of our jobs, because of something you just discovered on the internet?
I propose that your insights would be more welcome on a topic on which you are better-informed. If you want to shout desperately into the void about something you don’t understand, stick to Bitcoin.
Some Rules to Enfrce
Maybe it’s a blind spot (or, what’s the opposite of a blind spot?) I’m frequently mystified by the adoration for output from ChatGPT or Dall-E. How could someone see this as a replacement for a professional designer?
Then I remember the horse.
A “writer” produces writing, but that’s not all. I have to decide what to write. I have to ask and answer questions — before I start, while I’m working, and after I’m finished. I research a topic, I determine what my point of view is, and I decide which sources I want to rely on. I have to locate the window I was writing in, it was just here a minute ago.
Designers are are more than their output. They also ask questions: what’s this for, who are the customers, how do we want them to think or feel? Hey, what the fuck font is this? Can I use your Spotify login?
Do not reduce your understanding of someone to an example of the person’s output. You’ll get fooled or misled, as people frequently are technology they don’t understand.
It’s like trying to understand medicine by studying a tooth, and then declaring that a dentist can be replaced by a 3D printer.
In his remarkable article on ChatGPT, Stephen Wolfram describes how it generates convincing text that feels very human and lures us to imagine an intelligent author.
Wolfram then shatters that illusion by dissecting what the A.I. knows as it starts, and how it thinks as it works.
He comes suddenly to this realization — something that every writer, in some sense, already knows: it’s much easier to write an essay than it is to decide what it’s about.
Here’s Wolfram:
In the past there were plenty of tasks—including writing essays—that we’ve assumed were somehow “fundamentally too hard” for computers. And now that we see them done by the likes of ChatGPT we tend to suddenly think that computers must have become vastly more powerful….But this isn’t the right conclusion to draw…instead what we should conclude is that tasks—like writing essays…are actually in some sense computationally easier than we thought.
Totally based, Stephen Wolfram.
Not all Jobs can be Outsourced
The signs are everywhere. We’re entering a dynamic, volatile, and potentially dangerous time. Be warned: none of us are safe from stupid predictions.
Human workers still have some work left to do at the office. As long as there are two people left in the company, we’ll need to work together. This will require emotionally intelligent people like you and me.
We spend our days, and sometimes nights, asking questions, deliberating the answers, collaborating, negotiating, making decisions, crying in the bathroom, managing people, managing each other, and trying to figure out who to eat lunch with.
A.I. is no good at any of this stuff. Doing it well requires emotional intelligence or “EQ.” That’s the very opposite of what you get from A.I.
If we look back on history, it can seem like technology destroyed everything. But if we look more carefully, it might seem that it created us all.
A.I. can’t take any jobs we don’t give it. It can’t take away our ability to work with one another, to solve problems, to lead and inspire young people, to know and love one another, to see around corners, and to pick challenges to overcome that help to make the world look more like a place we want to live and work.
Human and humane leadership is also a prerequisite for a workplace where humans and machines can work well together. We need empathetic, respectful leaders to ensure that A.I. is designed responsibly, with respect for original content creators. They’ll help us develop safeguards to ensure the tech is used in safe and ethical ways, and to make sure that the tools aren’t used to exploit unsophisticated people.
No Remorse for The Thief & The Horse
If you think people skills are on the way out as A.I. is on the way up, you are wrong. Emotional intelligence has never been more important than it is at this moment. It is the tool we need to help us communicate and understand how we want to use this technology in our work and in our lives.
Crucially, this is a moment to protect and elevate our most inspiring and talented people. The youngest generation of future leaders should be encouraged to stretch towards opportunities to build, create, lead, and inspire.
I’m sickened by the idea that we’re sending the message that the scope of the opportunity in tech has somehow been reduced by this sloppy, silly commentary on A.I. We could be draining away meaning and purpose from our future in exchange for nothing more than a meaningless punch line.
That would be truly foolish.
It reminds me of another fable. This one is also about a talking horse. A thief is caught in the act, captured, and hauled before the King and Queen. They confer and order him to be summarily executed. The thief, thinking quickly, says “Queen, I have a mystical ability and a spiritual affinity with your prized Persian horse! Give me just one year, and I will teach your favorite horse to talk. If I fail, execute me.”
The royals confer again, they shrug, and agree to the thief’s bargain. As the guards lead the thief away towards the stables, another prisoner leans over and whispers “Fool! You’ll never teach the horse to talk. You’ll be executed for sure!”
“Just wait, who knows,” the thief replied. “So much can happen in a year!”
My old boss Scott Belsky at Adobe frames A.I. this way: a tool to be crafted responsibly and used to enhance the productivity of talented creators. He makes it look easy to talk about this subject in just the way I think it ought to be.