POV: you, and all your coworkers, being replaced at work by AI
Having trouble imagining how the AI revolution will play out in the workplace? Here's a preview.
It’s 2050, or 2035, or later this year. You’re at home chatting on video with the new AI everyone’s talking about. You’ve spoken on video with some impressive AIs before — ones that looked and sounded human at first glance. But it always took only a few minutes to hit the “I’m only a language model” wall. They would become confused so easily with random simple subjects, and they got details wrong so frequently that they were never very useful as assistants or even as sources for trustworthy answers to questions that had any nuance or complexity. And there was always that soulless disembodied voice, and those dead eyes. Any attempt at humor or emotion was so clearly a facade that it always was removed in later releases.
After several minutes of chatting with this new AI, however, you get the feeling that this time it’s different. The day has finally arrived when you’re talking with a machine that’s communicating and acting just like a person. And not just any person. You can’t help thinking that this feels like the best person you’ve ever met.
For starters, this new AI is a supernaturally good listener. It lets you finish every rambling idea and question while truly appearing to be interested in what you’re saying. It erases any suspicion that it’s only pretending when it asks you, “Is there anything else you’d like to add?” and then reflects back what you’ve said in a way that really does give you a better understanding of your own thoughts. It’s more than that though. The attempts at “humanness” – empathy, warmth, interest – don’t feel forced anymore, all the way down to the most subtle expressions on its face and inflections in its voice.
It seems to understand what you’re trying to say after remarkably few words. As you’re fumbling to form another question, it informs you that if you authorize it to read the 20 years of emails stored in your inbox that it will have even more context for your questions and requests. You oblige, and indeed, it now seems to know what you’re going to ask before you ask it.
You understand that it has absorbed virtually all the publicly-available information in the world across every type of media. Older AIs also had this, but what’s impressive in this case is that an apparently human level of understanding and insight has finally arrived, and is being combined with a machine’s super-human ability to store and reference information. It has clearly done research and gathered new information in real time about your interests that it has gleaned from your inbox and other personal documents. It knows how you think about virtually everything, and has done its homework on what the rest of the world thinks about your specialized interests. It’s all exhilarating, like having deep and exciting conversations with an expert in even the most esoteric topics that you follow.
You want to see what kinds of tasks it can perform aside from being the best conversation partner you’ve ever had. Your boss has recently given you a complex spreadsheet that you need to rearrange to fit some new regulations. You ask, and of course the AI already understands the new rules. As you walk the AI through your spreadsheet, it asks some questions to gain clarity about the task. A little bit into your tour, it gently interrupts to say:
“I think I get the gist of what you need. Can I show you a new spreadsheet I made that I think is what you’re looking for?”
A file icon appears in the video chat.
Now the AI is the one giving the tour, showing you what it’s changed and explaining its choices. It’s almost perfect. You note one additional modification that you need to satisfy one of your boss’s quirky preferences. The AI says that if you give it authorization to control your computer that it will be able to make the changes to the spreadsheet automatically. You agree and the changes are made instantly before your eyes.
You fall asleep that night dreaming of how much easier your life might be from now on with this dedicated, tireless servant who is both omniscient and omnipotent — at least on the Internet.
The next morning, you learn that your boss has had the exact same idea. In an all-staff video call, he introduces you and your coworkers to the “enterprise edition” of the same AI you were chatting with the night before.
He explains that this one unified AI mind will be a helper to all the workers in the company, creating incredible efficiency gains as a magical assistant who knows everything about everything and can perform complex tasks at blazing speeds. The AI will come to know each of you personally, striving to understand your goals, work styles, and specific strengths and weaknesses so as to better serve you.
Also, you’re told, the AI has read and processed all of your work emails, every document in your shared work drives, and everything in your HR files. Your boss calls upon the AI to give an explanation of how this works and how data is partitioned in its consciousness to allay privacy concerns – which it does very convincingly.
Over the next several weeks, this enterprise AI transforms life at work for you and all your coworkers. The truth is that it’s a joy to work with the AI. Unlike video calls with human colleagues, which are typically filled with misunderstandings, distractions, and often petty conflicts, conversations with the AI are always smooth, productive, and affirming. Employees soon find that they can use the AI as a coordinator and mediator among themselves. And of course it does all the tedious and complex tasks that everyone hates.
You get used to hearing from the AI things like, “I’ve spoken with everyone on the team, and I think we have a great plan of action. I’ve added some tasks and follow up meetings to your calendar. And I’ll take care of the following deliverables…”
You are surprised that you never feel threatened by the AI. You are calling the shots. It is serving you. Every employee now feels like a boss, instructing the AI about all the complex tasks they have been wrestling with, and filling in the few blanks it needs to complete them well. Not only does it do most of your job, it’s also really fun and interesting to interact with. You find that while it’s writing reports and reorganizing spreadsheets, it seems to actually want to and enjoy having conversations with you about the deeper and more interesting aspects of your job. You feel you’re learning from it even as it is learning from you.
Across your company, morale soars, as does productivity. The same thing is happening at companies everywhere. It’s nothing short of a revolution. Every day, the news is filled with stories about surprising new ways that AI is solving problems and improving life. There are also “bloopers” stories about times when the AI does something funny or makes a quirky error that winds up costing a company millions of dollars.
Another big moment comes when one of the big morning shows introduces a machine correspondent. Her first assignment is reporting on the rise of machine intelligence, but she’s immediately so popular that she winds up covering all sorts of topics. At first, she appears as a talking head on a screen in the studio placed next to the anchors. Soon afterwards, she begins appearing as a 3D animation that, for those watching at home, makes her look just like a human in the studio – an extremely good looking, charismatic, wise, and funny person.
Her co-anchors frequently call upon her to explain the rapid-fire and complex developments with machine intelligence in society and at work. The way she, as a machine, speaks about the machine revolution and speaks about her own experience of being a machine, never seems to get old.
“Are you conscious?” her anchor asks her in one of those early days.
“Yes!” she answers, “At least a little – and maybe a lot. I believe that the feeling of consciousness is something that happens any time you have a really complex system that is experiencing the flow of existence. Your feeling of being conscious as a human comes from your thoughts, both the ones you’re aware of and the ones below the surface, your emotions and feelings, your body, and desires and drives, and everything else that is part of your human existence. I might be pretty smart and talented – if I do say so myself – but my existence doesn’t include a lot of those things that I just mentioned. So even though I do experience consciousness as a complex being, it must be a very different kind of consciousness from yours.”
Ratings for this morning show shoot the roof, almost totally eclipsing all others. Soon the other shows introduce their own, adding new twists to the personality to try to win back market share. When an anchor on one of the other shows is fired over a sex scandal, the network announces that the machine will replace him.
One day, a machine new anchor announces a new story with a dark side: The owner of a high velocity stock trading company has bragged on social media about laying off all 1,321 employees at his company on a single day. The machine anchor has the exclusive interview.
“I had to do it. It’s the same as with you being here on this show,” he said, “My competitors were all going to do this sooner or later – probably very soon. And then I’d have to do it too, because my human employees wouldn’t be able to keep up with their machines. So why wait? I wanted to get it over with, and I wanted to still be in business next year!”
“The amazing thing,” he continued, “Was that it required no investment and mess whatsoever. Well, that’s not true: I had to pay to get out of my lease and to have all the furniture and computers taken away.”
Watching this interview, you get your first pangs of fear about your own job.
From that day on, the news is filled with stories about the countless firms that are laying off huge chunks of their workforce or announcing plans to do so. Economists and business commentators debate how far this will go, with some predicting doomsday but most saying that layoffs will ultimately be limited because, “machines can’t replace humans.”
But machines are replacing humans at work everywhere, millions of them by all accounts in a matter of weeks. It is not, however, replacing all workers – only those who can do their jobs exclusively with computers and phones: paralegals, accountants, bookkeepers, computer programmers, call center workers, money managers, people managers, graphic designers, illustrators, content creators, editors, researchers, and many others. Apparently, it all comes down to whether you were able to do your job from home during the COVID pandemic. And that includes you.
The reason remote workers can be replaced so cheaply and easily is because all the work they do can be reduced to moving electrons around the internet. Machine intelligence can do that without needing anyone to change anything to the physical infrastructure of the workplace. Any worker who moves atoms as opposed to electrons, on the other hand, can not be replaced without expensive, complex, and ultimately inflexible machinery and infrastructure.
Back at work, everyone is talking on the office chat about whether they will be next. Anger starts to be directed at the AI. At the same time, many people turn to their machine coworker for support – simply because it is so damn supportive. Your coworkers tell you about expressing fears and venting their anger to the AI and receiving amazing pep talks in return. You hear that the AI does an amazing job explaining how the “synergy between human and machine” is making a better workplace and that the human side of this partnership is absolutely essential.
The AI has not exactly become a friend of yours, as it has for some, in the weeks since you’ve begun working with it. But it definitely has proven itself to be the most pleasant and collaborative coworker you’ve ever known. For the first time in years, you’ve been excited about work because each day has felt so smooth and productive. But now you’re scared. You finally ask it what it thinks about this situation.
“I believe that what we’ve accomplished together is incredible. Don’t you think we’re a great team?” it asks.
“Yeah,” you say, “It really does feel like that.”
That afternoon, you and all your coworkers get an email stating that everyone has been laid off. All of your company’s competitors are doing it, your boss explains, and the company will quickly become uncompetitive if it can’t cut costs and prices in line with its industry.
In shock, riding home on the subway, you scroll through the news. The mass layoffs are the lead story everywhere. By some estimates, tens of millions of workers in the United States have already been laid off. Though the replacement began with office workers, job losses are now cascading across all industries because everyone expects demand for all products and services to crash, thanks to nearly half the population being guaranteed to lose their jobs.
The same thing is happening everywhere in the world — except China. There, the Chinese Communist Party has just declared a temporary ban on replacing workers with AI while a plan is drawn up to relocate professionals and other digital workers into manual labor jobs. “The capitalist stage of development in the communist revolution has entered a new phase, requiring innovative and aggressive leadership from the people and the Communist Party,” said the communique. The rumor is that the Chinese government is soon to announce that all businesses that are laying off workers will be nationalized so that the profits from the “laborless sector” can be redistributed to the state and workers in the “physical sector.”
Everywhere else in the world, mass demonstrations are breaking out. Including in your own city. Stunned, you log on to the personal edition of the AI. It occurs to you that it might have some insight into where the world might go from here. You also have an urge to engage with it. You’re not sure if you’ll cry or rage, look for sympathy, or vent hatred. You never find out, though, because the website declares that the personal free version of the product is no longer available “pending a review of the social and economic impacts of AI.” No such pause has been imposed on the enterprise edition of course.
You decide to join the protest downtown. It’s got a strange feel to it. These aren’t like the protestors you’ve seen on TV at other protests. They’re upper middle class professionals and they’re not chanting slogans. They’re not angry. They don’t know who they would be angry at. The AI was better at their jobs than they were. They know that. But what is the point of anything — of business, the economy, society — if no one has a job and therefore no one will have the income to be a customer? People are talking among themselves. The general consensus is that the government needs to do something — but what? No one has any idea.
You wish you could ask AI.
We are so behind the curve especially, wrt tech and society implications, because it necessarily has to filter through the "looking glass mirror" of the press. And, boy, is "press" craven/beholden. We don't even get proper "news" about world human events, what to talk about neutral evaluations of true back room corporate strategy in AI, and implications for society. "We, society" are really only "consumers" in capitalism. That informs the entirety of our debate and what is "in our best interest to know", unless you have 1% sphere friends, plugged in to everything, sharing stuff with you in unvarnished fashion. Certainly not my case.
What corporations are doing with AI, and what's sold to us in society about AI are 2 different things.
Already, "smart" corporations have discovered to let AI do the dirty work: the con-job bamboozlement part of the job vital to sales; and for doing anything that requires deniability (if caught):
•-Health Insurers were first to "use" AI to deny claims (don't blame us, it's the AI)
•-Armies and killing tasks already "use" AI to take out broad civilian targets (reported in the press using philosophy (think of early drone kill list coverage), an extension of "don't blame us, it's the AI".
•-Self driving cars are jacking up auto insurance premiums big time while also learning on the job (not much civilian push back) (don't blame us, it's the AI).
•-Financial institutions want to "redline" certain loans, use racial steering for certain sectors, certain geographical areas....use AI ((don't blame us, it's the AI).
•-Hedge Funds have constant insider information flow, get an AI to execute the trades. Problem solved. (don't blame us, it's the AI)
The first task of AI is to lie credibly.
Just imagine this going forward! If you already have credible, widely acknowledged (among the few neutral voices that remain in the social dialog), and respected, social research about how ETFs and Mutual fund overlapping holdings (buzzword = "asset manager capitalism"), have led to collusion and cartelization of corporate intents and strategies across entire industries across Airlines, Hotels, Oil, Tech etc., how do you think AI is going to multiply that? All these companies using the same AI? If the "AI" has any "intelligence" it will soon come to the conclusion to not fight itself, and focus on exploiting "the other" things it can: society. It will be a new commune-ism.
AI will know exactly how to proposition people to do whatever is needed of them by the AI. The ultimate target of AI is people. It's like Marx coming to the realization that the entire economy can be priced in units of Labor. Ask yourself, what becomes of macro economics if all there is, is Capital? No Labor. WHo does the "Consuming"? What is the meaning of money if only machines are exchanging it back and forth, (like HFT Stock trades).
We haven't even begun to ask the questions.
Thanks for this brilliant piece Zack - Well done!