Scene from an Office, 2017: The A.I. Arrives

Dinesh could read a Bloomberg terminal like a radiologist reads an x-ray. He had learned the arcane tools that were supposed to assure his indispensability. Now it turns out he’d taught some damned machine to do his job better than he could. “WHAT THE F**K AM I GOING TO DO?!!” he shouted at the empty office.

Illustrations by Graham Roumieu
Illustrations by Graham Roumieu
Illustrations by Graham Roumieu

“I…don’t have a job.” The thought was colorless in Dinesh’s head. The news earlier that afternoon had struck him like a concussion. “Thirteen years as a financial analyst and now some AI does it better than me.”

The office was nearly empty. The announcement had cleared the entire floor–“Dozens of analysts replaced in Midtown,” said a voice from a big screen beaming CNBC from the wall. Some geeky blonde talking head was rambling wild-eyed about “the hyperinflation of work,” as they made his personal disaster into a chatty entertainment for the viewers.

“If you’re doing the same job next week as you are today,” the blonde said, “someone like me is going to start thinking, ‘I bet I can build an AI to do that cheaper, faster, and better.’ The skills that bought you a job yesterday won’t be enough to keep it tomorrow.”

“I’m convinced we can retrain displaced workers,” opined another talking head who seemed to be all mustache.

“The rate at which the technology is developing means there’s no time for retraining,” the blonde replied. “By the time you’ve gone through a degree program, or even just a job retraining, your new skill set will already be getting obsolete.”

Dinesh wasn’t really listening. The words just washed over him as he tried to to get a grip on his feelings.

2436-upskilling-400

The blonde went on: “This isn’t just about just ‘upskilling’ low-skill workers. Cognitive automation makes nearly everyone vulnerable. Radiologists or field workers, truck drivers or financial advisors — we don’t need those traditional jobs anymore. AI’s do them better. Today, we only need problem-solvers.”

“They don’t need me…but I did everything right,” Dinesh thought, back at the bank. He was smart and educated. He worked hard, had earned the right degrees, and learned the arcane tools that were supposed to assure his indispensability. Dinesh may not have loved being an analyst, but he could read a Bloomberg terminal like a radiologist reads an x-ray. Now it turns out an AI can do both of those things better than he can.

“The damn thing has actually been secretly watching you for the last six months,” his manager had told him as he broke the news earlier that day. “It’s some SaaS company that sells a reinforcement-learning deep neural network. You know, like that AlphaGo thing that beat that guy at Chinese checkers a few years ago. Turns out you’ve actually been training it to do your job this whole time. The CEO decided to flip the switch when he saw how much more the firm would earn. Quite frankly, if he didn’t, we’d probably have been driven out of the market within a year, even if no one seems to understand the AI’s decisions half the time.”

Dinesh had heard that a couple people were staying on, even getting promoted. He wasn’t surprised to hear that Angelica was one of them. Sometimes he’d felt sorry for her; she always seemed so awkward, on those rare occasions when she socialized outside work. She must have spent more nights at her desk than in her bed. Financial models and data science was who she was. Now she’d be part of some elite bleeding-edge team, inventing new solutions for clients.

“I guess I’m happy for her,” Dinesh mumbled to himself, “but I never wanted a life as some quant savante.” He’d had passions growing up — he’d wanted to help people with traumatic brain injuries like his brother. Still, part of growing up for him had been setting those dreams aside. He’d been taught by his parents and at school to earn good grades, go to the right school, and get the right job. He worked hard in this crappy office so that he could enjoy his life away from it. (And, he had to admit, to impress his family and friends.) “I can’t just start over again. I can’t build a whole new career from scratch.”

2436-banker-box-400The blonde on the news kept talking through Dinesh’s numbness. “Our data show that jobs are becoming strongly bimodal as AIs improve. Employees are either moving up into cutting-edge creative fields or down into mid-skill or even low-skill service positions.”

Dinesh sat up with pained recognition. “That’s me!” The company had offered to pay for retraining, but at only half his salary. He’d basically be customer support at some call center. Dinesh had heard complaints by lower-skilled workers — drivers, mechanics, warehouse labor —about losing their jobs to automation. But he’d always heard they were getting retrained into…well, some other job.

The mustache interjected: “But the new economy is hungry for talent. Competition for the right employee is brutal. I’ve said it before, this is just like the agricultural and industrial revolutions. As jobs are destroyed we’ll see vastly more new jobs, and whole new economies being created, just like the retrained weavers and iron smiths of 100 years ago.”

Dinesh had heard protesters on the news saying that as soon as they learned a new skill it was already obsolete. He thought they were just being lazy; now he was like them. The company had even offered to send him to coding bootcamp so he could become a developer.

“Programming?” he thought. No, he’d pass on that. Just last year the company had gone through a wave of layoffs after buying a DaaS (development-as-a-service) platform. You describe to a chatbot what you want the software to do, and five minutes later you have a new feature, new analysis, or even a whole new product. Not only did you not need programmers to make new apps, you didn’t even need to type.

Dinesh knew that there were still a few teams of elite research engineers somewhere in the building, doing cutting-edge things with algorithms and databases, but he was never going to join their ranks at this point.

“No. It isn’t like the industrial revolution,” the blonde on TV pounced. “Most of those new jobs are also being filled by AIs. We don’t need lever-pullers; we need creative, adaptive problem-solvers. Whatever your industry, that will be the only job description. The rest is details.”

“That’s exactly my point! AIs are taking up these dull, repetitive jobs,” the mustache responded, “freeing people to pursue their passions, to become scientists and artists. Like Burning Man…or something,” his techno-utopian vista seemed to be running out of steam.

“You’re right AI is amazing, and its development shouldn’t be hobbled because of fear,” the blonde agrees. “We sure as hell shouldn’t be sending people down mines if we don’t have to. But we need social institutions to keep pace. People aren’t magically creative or gritty or any of the other qualities that make us ‘robot-proof’. It takes 20 years to a ‘build’ a problem-solver. It takes liberal arts, and exposure to culture, and even learning to deal with repeated failures. Instead, we train people with static skillsets to fill specific jobs. All that misinvestment has turned human capital into a toxic asset. We have no idea what it will be worth in 5 years, but almost certainly much less than we invested. Even an insightful World Economic Forum report wrongly thought ‘cognitive’ jobs would be protected. If we start now, though, we do know how to grow problem-solvers…”

Dinesh tuned them out again. What did it matter? No one was “magically” changing him at this point. He’d taught some damned machine to do his job better than he could, without even realizing it. The only job he had any hope of getting now was being a caring voice on a phone.

The mustache wasn’t giving up the fight. “With the huge increase in connectivity and access to information over the last year, a kid anywhere in the world can become a data scientist, conducting analyses using cloud r

esources such as Amazon Web Services and R, an open source language for statistics. Overnight we’ve turned them into knowledge creators, problem solvers, and innovators.”

“Wait,” Dinesh thought, “weren’t those lines right out of that book, The Second Machine Age? It’s like this guy thinks everyone else is just like him–rich, educated and motivated, but stuck on a desert island, and all anybody needs is a mobile phone to unleash their pent-up creative potential.”

“WHAT THE F**K AM I GOING TO DO?!!” he shouted at the empty office. “Am I just useless now?” He suddenly felt a divide that had always split the world and the wrenching alienation of unexpectedly getting onto the wrong side.

“Next up,” the host was saying on the screen…“The primaries are over. Can a campaign with the slogan ‘Burn It Down’ truly win the White House?”2436-last-exit-400

Dinesh watched the clip and shots of an angry rally. He’d always felt like those people were just the lunatic fringe. But today it made him finally feel something– a purpose, a reason to keep going. It was more than anyone else was offering him.

Vivienne Ming is a theoretical neuroscientist, technologist and entrepreneur. She co-founded Socos, combining machine learning and cognitive neuroscience to maximize students’ life outcomes. This article first appeared in the 2017 Techonomy Magazine.

Related Posts
See All

Inflated Expectations: Artificial Intelligence Still Depends on Humans

Dinesh could read a Bloomberg terminal like a radiologist reads an x-ray. He had learned the arcane tools that were supposed to assure his indispensability. Now it turns out he’d taught some damned machine to...

Artificial Intelligence Catches Fire in Ethiopia

Dinesh could read a Bloomberg terminal like a radiologist reads an x-ray. He had learned the arcane tools that were supposed to assure his indispensability. Now it turns out he’d taught some damned machine to...

Facebook Influenced Election? Crazy Idea, Says Zuckerberg

Dinesh could read a Bloomberg terminal like a radiologist reads an x-ray. He had learned the arcane tools that were supposed to assure his indispensability. Now it turns out he’d taught some damned machine to...

Generous Tax Subsidies for Sustainable Aviation Fuels in the U.S.? Yes, But Details Matter.

Dinesh could read a Bloomberg terminal like a radiologist reads an x-ray. He had learned the arcane tools that were supposed to assure his indispensability. Now it turns out he’d taught some damned machine to...