By Conor Sen
These days, there are various publications, posts and videos warning of armies of robots preparing to flood jobs in the fields of technology, finance, marketing and wherever else imaginable. And there is no doubt that artificial intelligence is rapidly changing the way we live and work. In the midst of all this, however, there has been a relative calm in the labour market, which is expected to last at least until the end of the year.
For workers, there were two fears about 2026. First, that the continued slowdown in job growth would push the unemployment rate higher for the fourth year in a row, and second, that rapid advances in artificial intelligence would accelerate job losses. On both fronts, the reasons for concern now seem less.
Labour market data is arguably more stable today than at any time since the Federal Reserve began raising interest rates in 2022 to curb rising inflation. The unemployment rate has returned to last summer's levels, having fallen slightly since the government shutdown late last year. Weekly claims for unemployment benefits are at very low levels, despite economic and stock market turbulence due to the war in Iran, and the number of people who continue to receive unemployment benefits is no longer increasing compared to a year ago. As Fed Chair Christopher Waller noted last week, there is now a better understanding of limited labour force growth, which means that not many hires are required to maintain the balance.
However, concern about AI has, if anything, intensified this year, and we see it in the headlines as well. Two months ago, a report by Citrini Research shook markets as it presented a scenario in which AI would lead to massive job losses within 18 months. Last week, Anthropic CEO Dario Amodei said he was concerned that half of the entry-level positions in sectors such as technology, finance and consulting, although initially increased, would be replaced by AI systems within one to five years.
It is worth noting the explosive success of Anthropic's Claude services, which helped increase the company's revenue from $9 billion to $30 billion by the end of 2025. Such growth portends significant changes for employees in the near future.
However, the rapid growth of AI brings challenges, both for companies like Anthropic and for the customers who use their products, which "give away" time to employees before we see any disruptions to employment.
It is unclear, for example, whether Anthropic has the resources needed to manage the growing demand for its products in the short term. The company recently decided to charge large users based on usage rather than a fixed amount, as compute-intensive products such as Claude Code are rapidly increasing costs. There is also a growing number of complaints about the performance of the models and their reliability, as demand has skyrocketed. At the same time, the Wall Street Journal spoke of a "sharp capacity crisis" in the IT sector that could limit the usefulness of new AI tools.
"Computing power has emerged as the key limiting factor in the expansion of AI," Goldman Sachs analysts noted last week.
On the customer side, AI is now an increasingly important piece of companies' budgets. "Tokenmaxxing" — or the overuse of tools like Claude, OpenAI's ChatGPT and Alphabet's Gemini — by employees eager to prove they're not tech "dinosaurs" is gaining traction, according to Kevin Roos of the New York Times. However, tokenmaxxing and Anthropic's rapid revenue growth mean that companies are rapidly increasing their spending on AI, which will soon become large enough to require greater scrutiny from management and investors.
As costs rise, companies are expected to start trying to better understand how productive these tools are and become more selective about how they are implemented. What confuses things even more is the "AI-washing" by executives who prefer to blame technology for any job cuts, regardless of the actual reasons, because this is seen as better in the eyes of investors.
Wells Fargo economists find that the difficulties associated with AI are focused on new employees, while opportunities for more experienced workers either remain unchanged or have improved. "AI appears to be reshaping the composition of labour demand, rather than unequivocally reducing it," a group led by economist Tom Porcelli wrote this week.
If workers are worried about what the development of AI will mean for their jobs, it's not the boom that should concern them, but the recession. Job losses due to technological disruption tend to occur in times of recession rather than in a linear fashion. Ernie Tedesi, chief economist at Stripe LLC and a former Bloomberg Opinion columnist, used the example of travel agents. In this case, employment declined rapidly only during and after the 2001 recession, and not during the internet boom of the late 1990s.
The risks associated with AI are likely similar. Corporate earnings remain high, and while some companies are cutting back on spending to fund their AI investments, their revenues continue to grow. The pressures on the labour market associated with declining revenues, earnings, and stock prices seem distant for now. In the short term, the lack of computing power will also limit the pace at which more sophisticated models can be adopted.
For workers, this year may look like the eye of a storm: a brief respite before the oncoming storm.
BloombergOpinion
