Author: Neil Harrington
Perusing business and economics sections of national and local newspapers recently might lead some observers to think the tech industry is collapsing. Google cut 12,000 jobs from its ranks, or about 6 percent of its workforce. Amazon announced plans to lay off 18,000 people. And it’s not just the giants; smaller tech companies such as SoundHound laid off 200 workers, half of its workforce. The tech job tracker layoffs.fyi shows almost 250,000 tech jobs have been lost since the start of 2022. With new waves of layoffs seemingly announced every day, the headlines have been omnipresent, especially for workers in the tech sector unsure if they will have a job when they wake up.
So, what’s causing all these job losses in tech? Overinvestment during a long period of cheap money, not AI, appears largely to blame. However, it’s hard to fault anyone for thinking about recent advancements in AI, specifically the launch of ChatGPT, and envisioning a major disruption brewing. If someone is the first to see an AI revolution coming, it would be tech companies. As ChatGPT, Bard, Ernie Bot, and other AI products continue to develop, they could increasingly disrupt labor markets, or slowly shift tasks for workers within occupations, or quietly fade in people’s memories as other tech advancements have done. It is still far too early to know exactly how and if these technologies will disrupt labor markets, and the full extent of disruption won’t be clear until it shows up in labor market data. Whatever happens, LEAD is closely monitoring trends and will update analyses as disruption happens (or not).
AI didn’t cause recent mass layoffs in tech…
Most of the news coverage around tech layoffs note the sector’s overinvestment over the past several years. An extended period of low interest rates leading up to and during the first part of the pandemic made accessing capital virtually free for companies. This fueled soaring profit margins for tech companies and their investors and accelerated employment growth.
For example, in 2021, the tech giants Facebook and Amazon had more than 5.5 times the number of employees they had in 2015, while Google and Apple’s workforces expanded by a still lofty 153 and 40 percent, respectively, according to the investor research platform Macrotrends. Similarly, smaller tech firms such as Salesforce and Carvana saw employment boom while money has been cheap. But as macroeconomic policy has sought to fight inflation through higher interest rates, investments have slowed, and tech firms are finding it harder to maintain large workforces with high salaries.
…but AI is and will continue disrupting tech and other industries
An ominous message spotted in the halls of a DC consulting firm.
Overinvestment and increased borrowing costs seem to be the main cause for the current wave of tech layoffs, but job cuts are also happening as these companies grapple with the disruption of AI advancements such as ChatGPT. Released at the end of 2022, ChatGPT is an AI language model developed by OpenAI that can generate human-like text based on patterns learned from vast amounts of internet text data. ChatGPT responds to simple requests like “explain what ChatGPT is in one sentence to someone not familiar with machine learning” and returns text like the last sentence or offers more detailed responses to more sophisticated prompts. The UNC Kenan Institute’s Five Economic Trends to Expect in 2023 even features an economist-sounding paragraph generated by ChatGPT on how AI will impact workers.
After its release, major tech companies have scrambled. Google reportedly issued a ‘code red’ over the threat ChatGPT poses to its search engine business. Resultingly, the company’s executives restructured teams and directed many workers to devote more time to AI products and tools, but ultimately flubbed the initial release of Bard, its ChatGPT rival, leading to a 10 percent loss in its market value. Even if AI isn’t being blamed for recent layoffs, current and future AI products and advancements are likely weighing on companies’ projections of future labor needs and expected productivity.
ChatGPT, AI, and the future of work
Of course, ChatGPT has its limitations and has not caused the thousands of job losses in tech. However, as a language model that writes decently well, it is likely to increase the average writing ability of workers and perhaps increase productivity for people with writing-heavy jobs or others who find innovative uses. CNET and other publishers have even started using ChatGPT to write articles and generate poll questions and quizzes.
The labor market disruption potential of ChatGPT and other AI advancements will largely depend on how they are used. This has been the case with previous tech advances as well. It took nearly 50 years for electricity to revolutionize manufacturing in a way that made productivity soar, largely because business owners were hesitant to move away from steam power. Map applications on smart phones first allowed taxi drivers to find more efficient routes before giving rise to the likes of Uber and Lyft, which in turn severely disrupted the taxi labor market. Still, recent technological progress has had relatively little impact on productivity and labor markets, as evidenced by the past few decades of huge tech advancements but lackluster productivity growth. As new versions of language models and other AI are released, people and businesses will find innovative uses which could alter labor markets, with impacts perhaps lurking beneath the surface.
Making sense of ChatGPT’s and AI’s impact
In short, ChatGPT and AI advancements will undoubtedly disrupt certain jobs and the tasks that are performed in some way—the severity of those disruptions will become clearer as it happens and shows up in labor market data. For example, most estimates of future employment expect technological progress to have a greater impact on production industries and non-routine occupations because that’s what historical data indicates will happen.
What makes AI potentially different, or alarming to some, is the potential pace of change and the types of work it may affect. As a technology that learns on its own and improves with the amount of interaction it has with humans, AI advancements could come much quicker than other technologies. And for the first time, ChatGPT has suggested that technology disruption may significantly impact analytical and creative jobs as well. But disruption does not necessarily mean destruction.
At this point, it appears that disruptive AI will impact North Carolina eventually, as it will the rest of the world. However, it’s still far too early to fully understand the scope of that impact. If history is a guide, it’s more likely to change how we work (for the better) and create new occupations than to decrease the overall demand for human labor. For now, LEAD partners can take solace in the fact that we’re closely monitoring trends, using a variety of long-term and real-time sources, and are prepared to help people deal with the uncertainty surrounding automation and AI advancements. A couple years from now, we may all be celebrating the loss of lower paid, routine work and reaping the benefits of better-paying, more challenging and rewarding employment spurred by advancements in AI.