When ChatGPT came out last November, Olivia Lipkin, a 25-year-old copywriter from San Francisco, didn't think much of it. Then articles about how to use the chatbot in the workplace appeared in the tech start-up's internal Slack groups, where she worked as the company's sole author.
Over the next few months, Lipkin's duties shrunk. Managers began referring to her as “Olivia/ChatGPT” on Slack. She was fired in April without explanation, but when she saw managers writing about how using ChatGPT was cheaper than paying a writer, the reason for her firing seemed clear.
“Whenever people mentioned ChatGPT, I felt insecure and worried that it would replace me,” she said. “Now I actually had proof that it was true, that those fears were justified, and now I was actually unemployed because of AI.”
Some economists predict that artificial intelligence technologies like ChatGPT could replace hundreds of millions of jobs, leading to a catastrophic workforce restructuring that mirrors the Industrial Revolution.
For some workers, these effects are already being felt. Those who write marketing and social media content are among the first wave of humans to be replaced by tools like chatbots that seem capable of creating plausible alternatives to their work.
Experts say even advanced AI can't match the writing skills of a human: it lacks personal voice and style, and often provides incorrect, nonsensical or biased answers. But for many companies, cutting costs is worth sacrificing quality.
“We really are in a crisis,” said Sarah T. Roberts, an associate professor at the University of California, Los Angeles, who specializes in the digital workforce. “[AI] comes for the jobs that should actually be automation-proof.”
The quality of artificial intelligence has increased rapidly over the past year, spawning chatbots that can carry on fluid conversations, write songs and produce computer code. In an effort to make the technology mainstream, Silicon Valley companies are spreading these products to millions of users, often offering them for free — for now.
AI and algorithms have been part of the world of work for decades. For years, consumer goods companies, grocery stores and warehouse logistics companies have used predictive algorithms and robots with AI-powered vision systems to make business decisions, automate some routine tasks and manage inventory. For much of the 20th century, industrial plants and factories were dominated by robots and countless office tasks were replaced by software.
But the latest wave of generative artificial intelligence — which uses complex algorithms trained on billions of words and images from the open internet to produce text, images and audio — holds the potential for a new level of disruption. Experts say the technology's ability to produce human-sounding prose is putting high-paid knowledge workers in the crosshairs for replacement.
Reporter Danielle Abril tests columnist Geoffrey A. Fowler to see if he can tell the difference between an email written by her and ChatGPT. (Video: Monica Rodman/The Washington Post)
“Every previous automation threat has been about automating the difficult, dirty, repetitive tasks,” said Ethan Mollick, an associate professor at the Wharton School of Business at the University of Pennsylvania. “This time, the threat of automation is aimed squarely at the highest-earning and most creative jobs that…require the highest educational background.”
In March, Goldman Sachs predicted that 18 percent of the world's work could be automated by AI, with white-collar workers such as lawyers at higher risk than those in jobs such as construction or maintenance. “Occupations that require a significant amount of time spent outdoors or performing physical labor cannot be automated by AI,” the report says.
The White House also sounded the alarm, saying in a December report that “AI has the potential to automate 'non-routine' tasks, exposing large new segments of the workforce to potential disruption.”
But Mollick said it's still too early to estimate how disruptive AI will be to the workforce. He noted that jobs such as copywriting, document translation and transcription, and paralegal work were particularly at risk because they involved tasks that could easily be completed by chatbots. High-quality legal analysis, creative writing or art may not be as easy to replace, he said, because humans still outperform AI in those areas.
“In general, think of AI as acting like a high-end intern,” he said. “Jobs that are mostly designed as entry-level jobs to get you into a field where you do something meaningful, but they are also kind of a stepping stone to the next level – these are jobs that are under threat.”
Eric Fein ran his content writing company for 10 years, charging $60 an hour to write everything from 150-word descriptions for bathmats to website copy for cannabis companies. The 34-year-old from Bloomingdale, Illinois, built a stable business with 10 current contracts that provided half of his annual income and provided a comfortable life for his wife and two-year-old son.
But in March, Fein received a message from his largest customer: his services would no longer be needed as the company was switching to ChatGPT. Fein's nine other contracts were gradually terminated for the same reason. His entire copywriting business disappeared almost overnight.
“It wiped me out,” Fein said. He urged his customers to rethink and warned that ChatGPT cannot write content with its level of creativity, technical precision and originality. He said his customers understood that, but they told him that using ChatGPT was far cheaper than paying him his hourly wage.
Fein was rehired by one of his clients who was dissatisfied with ChatGPT's work. But it's not enough to support him and his family, who still need to stay financially afloat for just over six months before they run out of money.
Now Fein has decided to take a job that AI can't do and has enrolled in courses to become an HVAC technician. Next year he wants to train as a plumber.
“A trade is more future-proof,” he said.
Companies that replaced their employees with chatbots faced major problems. When technology news site CNET used artificial intelligence to write articles, the results were riddled with errors and led to lengthy corrections. A lawyer who relied on ChatGPT for a legal assignment cited numerous fictitious cases. And the National Eating Disorders Association, which fired employees from its hotline and reportedly replaced them with a chatbot, stopped using the technology after it distributed insensitive and harmful advice.
Roberts said chatbots can cause costly errors and that companies rushing to integrate ChatGPT into operations are being “hasty.” Because they predict the statistically most likely word in a sentence, they intentionally produce average content. That presents companies with a difficult decision, she said: quality vs. cost.
“We have to ask ourselves: Is a facsimile good enough? Is imitation good enough? Is that all we care about?” she said. “We will lower the level of quality, and for what purpose? So that the company owners and shareholders can get a bigger piece of the pie?”
Lipkin, the copywriter who found out she was replaced by ChatGPT, is completely rethinking her office work. She initially started content marketing as a way to make a living while pursuing her own creative writing. But she found that the job burned her out and made it difficult for her to write herself. Now she is starting a job as a dog handler.
“I’m taking a complete break from the office world,” Lipkin said. “People are looking for the cheapest solution, and that’s not a human, it’s a robot.”
