GenAI's job threat turns up in job ads, not layoffs
Adoption of generative AI is akin to offshore outsourcing, where low-level tasks are shifted elsewhere. Instead of work going to India, it's now going to the machine.
Generative AI is reshaping the job market, subtly influencing job postings rather than triggering widespread layoffs. In sectors prone to automation, labor market experts are seeing evidence of a noticeable shift toward hiring fewer employees in occupations most at-risk to GenAI's automation capabilities.
Despite the tech industry experiencing more than 33,000 layoffs this year, the job cuts are seldom attributed directly to advancements in automation. Employers are still navigating the initial phases of integrating large language models into their operations. ChatGPT-3, the frontrunner in touching off the GenAI revolution, has been around for only just over a year.
However, the potential for GenAI to disrupt the labor market in major ways remains.
GenAI is a new form of offshore outsourcing, but instead of jobs going to India or China, this time they're going "in the machine," said Deborah Compeau, an information systems professor and interim dean of Washington State University's Carson College of Business and Research.
Compeau said that GenAI like offshore outsourcing is taking on low-level work while remaining jobs require higher levels of learning. For instance, GenAI can "produce strategic plans that are copies of things that existed before," but she said some companies will also want employees who can develop novel approaches.
GenAI and jobs
GenAI has the potential to affect a broad swath of white-collar work, including clerical, finance, legal, computer and mathematics jobs, according to a new study, "Generative Artificial Intelligence and the Workforce." The project was a collaborative effort between The Burning Glass Institute and the Society for Human Resource Management.
GenAI's first effect will be disruption, followed by job expansion as productivity increases, said Gad Levanon, chief economist at The Burning Glass Institute. For example, if software developers become 50% more productive due to AI, a company may not need as many developers to complete its projects, even if demand grows by 20%. "You still have an overcapacity," he said.
With software developers, GenAI may increase demand for developers. But it's difficult to know what the final effect of two competing forces -- job disruption and job creation -- will be on those developers, Levanon said.
For other occupations, including clerical, finance and insurance, GenAI will lead to a net decline in the number of jobs, Levanon said. The effects will be gradual -- not likely through sudden mass unemployment but through changes in hiring. "They will stop hiring new people for those positions," he said. That may already be happening.
Occupations that are more exposed to AI had a bigger drop in online job ads in 2023.
Gad LevanonChief economist, The Burning Glass Institute
The Burning Glass Institute analyzed occupations most exposed to GenAI and then at the growth in online job advertisements by occupation for 2022 versus 2023, finding what may signal AI's impact.
"Occupations that are more exposed to AI had a bigger drop in online job ads in 2023," Levanon said.
Victor Janulaitis, CEO at Janco Associates Inc., a labor market research firm, said that in IT, employers are turning to chatbots and AI tools that generate code for some entry-level tasks, such as staffing help desks. He added that companies are now typically recruiting for experienced individuals rather than entry-level candidates.
Janulaitis said he could find job ads for someone who is a blockchain manager, but "I can't find job ads for someone who is an entry-level blockchain programmer," he said.
AI and software development
Flavio Villanustre, vice president of technology and global chief information security officer at LexisNexis Risk Solutions, said GenAI is not the primary cause of the recent tech layoffs.
"Executives and investors seem to be more focused on profitability and efficiency rather than technological moonshots," Villanustre said. Nonetheless, GenAI efficiency gains "could have an impact on the number of software engineers required in some, or most, large technology organizations."
If use of systems such as GitHub Copilot or Amazon CodeWhisperer can achieve 10% efficiencies in software development teams, it could mean 10% of those software development positions could be made redundant, Villanustre said.
A report from Washington State University's Carson College on AI and business readiness found that more than half of the 1,200 full-time working professionals it surveyed now use AI in their jobs in areas such as data analysis and content generation. However, there is significant concern among those surveyed that they may be left behind in their careers if they do not learn how to leverage AI.
Adjusting to these changes might mean, for instance, more "micro-credentials" that provide targeted training on specific areas that may require three classes, with a depth of learning that a university can provide, Washington State University's Compeau said. These can be both for degree holders and those without a degree.
Compeau said colleges also must rethink their models, "and like the business world, we are going to be racing to keep up."
Patrick Thibodeau covers HCM and ERP technologies for TechTarget Editorial. He's worked for more than two decades as an enterprise IT reporter.
