IT leaders face tension between AI and hiring

For IT leaders, a new conundrum is emerging: Do you hire specialized professionals, or burn through the IT budget on AI tokens? While some companies are replacing workers with AI, other leaders are deciding they need humans in seats to meet their IT goals.

According to executive outplacement firm Challenger, Gray & Christmas, AI is the most cited reason for the recent burst of tech layoffs, which hit 97,000 in May. One example is crypto exchange Coinbase, which laid off 14% of its staff at the beginning of May, reportedly in response to market conditions and the rise of AI. In an X post, Coinbase CEO Brian Armstrong described how the company is embracing “reduced pod sizes, including ‘one person teams’ with engineers, designers, and product managers all in one role.”

Meta also cited AI in its decision to reduce its workforce by approximately 10% in May, along with roughly 8,000 employees expected to see their roles eliminated as a result of the technology, NPR reported.

However, not all companies are embracing the idea of substituting staff for AI. According to data from consulting firm Robert Half, as reported by Fast Company, 32% of hiring managers said they’d rehired for a role that was previously cut due to AI and automation. Another report by Forrester suggested 55% of employers regretted cutting employees in favor of AI tools.

Tokenmaxxing or salarymaxxing? Nick Deveau, CEO and co-founder of a rental vacancy AI platform Grotto AI and Andrew Moore, CEO of AI software platform Lovelace.ai, suggested AI could limit future hiring rather than reducing current headcount.

Grotto AI has postponed some hiring because of in-house AI tool Greta, which helps the organization evaluate its valuable assets. Greta has the total context of the business and can perform different functions, Deveau explained, but can’t replace the discernment of human employees.

“[Greta] just helps us get to conviction more quickly around where we need to hire people, because I still think that hiring either experts or really high potential people is going to yield more return than just letting an AI do the job,” Deveau said.

Moore suggested he hasn’t needed to lay off anyone, despite his engineers’ roles changing radically because of AI-powered coding tools. He added that it’s hard to get an AI replacement right, especially given the costs of integrating automated tools into tech stacks and workflows.

Moore said one of the biggest problems faced by tech leaders is not being able to tell if the technology will result in higher- or lower-quality output, or how much labor will actually be saved.

“I’ve had to hire far fewer engineers than I was originally budgeting for,” Moore said. “All of us are doing about two to three times as much coding as we were before we were using these AI tools.”

Deveau also suggested AI tools aren’t strong enough yet to replace humans’ complex thinking, adding that the concept of tokenmaxxing seems like a “trap”: while it’s easy to adopt, it might not yield the best results.

“I’m much more worried about our teams offloading their strategic thinking to these sorts of tools today than I am about those tools getting so good at it that they’re going to displace [human employees],” Deveau said. “Let’s say you’re trying to make some really important strategic decision, and you just immediately start chatting with Claude, as opposed to spending what you would have used to have done, which is 30 minutes with a blank canvas.”

Where to go from here. Deveau said IT professionals should consider joining organizations where their skills are critical to growing the business and can be accelerated with AI.

Meanwhile, Moore emphasized the importance of AI-skilled engineers, especially as organizations attempt to use the technology to shrink teams. Many companies, including Lovelace.ai, “would never even dream at the moment of hiring an engineer who wasn’t already very familiar with how to make best use of automation and AI coding tools,” he said.