TLDR
- Most companies adopting AI are seeing strong productivity gains, but relatively few are using those gains to cut jobs, suggesting AI is driving reinvention rather than mass layoffs, according to Dan Diasio, EY global consulting AI leader.
- Workers feel excited yet anxious about AI, with many teaching themselves how to use new tools because employers invest more in technology than in training or mindset shifts.
- EY’s findings indicate that organizations focusing on growth and new capabilities – rather than cost cuts – are more likely to turn AI investments into competitive advantage.
Workers are conflicted about the impact of artificial intelligence in the office, expressing both enthusiasm about its productivity boosts while voicing anxiety about job security and uncertainty over how AI will reshape their roles, according to a recently released EY survey.
It doesn’t help that headlines about companies cutting thousands of jobs only reinforce worries about AI-fueled cost cuts. But the reality is much more nuanced, said Dan Diasio, EY global consulting AI leader, in an interview with The AI Innovator.
He said a lot of factors go into a decision to lay off workers, some due to AI, others due to things like economic uncertainty. “Just like any other big revolution, the jobs will evolve,” Diasio said. “There will be some work that is replaced by AI, and others will move to new jobs that start to amplify what is differentiated, what is creative with these systems.”
His point: If a company merely uses AI to reduce the workforce and let bots do the work, then it will resemble the next company that is doing the same. “You lose differentiation when everybody is using the same model,” Diasio said. Thus, cutting jobs to save money should not be the main goal of AI deployment, especially for forward-thinking companies.
Moreover, deploying AI mainly for cost cuts assumes the company wants to maintain the status quo instead of grow. “Once you use AI to be able to optimize a portion of work, then you should start to ask the question … ‘What new things can we do?’” Diasio said. “If it used to take me 45 hours to do the work, or a month to do the work, and now I can do it in an hour or in a day, what does that change about the way that I run my business?”
The latest EY U.S. AI Pulse Survey, released this week, supports his premise. It shows that 96% of organizations investing in AI are seeing productivity gains, with 57% calling those gains significant. Yet only 17% of respondents said those improvements translated into reduced headcount.
About half of the companies reported reinvesting AI-driven gains into existing AI capabilities, new product development, cybersecurity, R&D, and upskilling employees. The study queried 500 senior U.S. decision-makers.
Notably, the tech shift remains in the early innings. “Anytime there is a big technological revolution, it starts with fear and uncertainty and ends with some reinvention,” Diasio said. “We’re just in the early days where it is dominated by fear, and we haven’t come out the other end of the reinvention.”
Surprising workplace findings
One of the surprise findings was that employees were mainly left on their own to figure out how to use AI at work. More than 80% of workers reported teaching themselves how to use AI tools, rather than receiving structured training. “Many people still feel like they need to figure it all out themselves,” Diasio said. “They were teaching themselves how to do it.”
At the root of the problem, Diasio argues, is an imbalance in how companies invest in AI. Most spending and attention have gone into tools – buying licenses, deploying models and standing up platforms – while comparatively little effort has been devoted to mindset and skill set.
“If you break success down into three areas – let’s say mindset, skill set and tool set, most of the effort and most of the capital … is focused on the tool set,” he said. “They’re not investing yet at the same proportion in the mindset and skill set to be able to help people understand what creates differentiation, what does good adoption look like, and what are those human skills that we need to strengthen to be able to work better with AI.”
That imbalance can reinforce fear. When AI is framed primarily as a productivity lever, employees tend to assume savings will come from job cuts. Diasio said organizations that focus instead on growth – using AI to expand capabilities or enable new offerings – are more likely to channel productivity into competitive advantage rather than cost reduction.
EY’s AI Pulse supports that view: 56% of companies that reported positive returns on AI investments were more likely to increase their AI budgets. Organizations investing $10 million or more in AI were far more likely to report major productivity gains than those investing less (71% vs. 52%).
Another contrarian finding was that younger workers reported higher levels of anxiety about AI than older colleagues. New graduates face a tighter job market, and productivity-driven automation often targets tasks typically assigned to less experienced employees.
“To be a good ‘context engineer,’ to be able to work well with AI, you need to have good creativity and a good depth of knowledge and expertise,” Diasio said. “That expertise … has been built up by experience.”
To stay relevant, Diasio emphasized three skills workers can start developing immediately: multidisciplinary thinking, systems or critical thinking, and advanced communication skills. AI systems are controlled almost entirely through language, he noted, and the ability to use nuanced instructions is often what separates average output from differentiated results.
In one example, he described AI filmmakers who used the same tools available to anyone but were able to create a more sophisticated movie by experimenting with tens of thousands of prompt variations, effectively learning how language serves as a control panel for complex systems. “They learned how to use language as a control panel to be able to control all these dials in the output,” Diasio said.
Looking ahead, will the workplace of the future be led by humans dictating what they want AI to do? Or will it be led by AI, which will do most of the work and escalate an issue to a human only when it’s needed?
Diasio is betting on a human-led workplace. In his view, letting machines dictate the critical path of work risks pushing companies toward “statistical sameness,” where businesses converge on similar outputs and lose differentiation. “I believe much more in the former, where it’s people who are going to be directing AI,” he said. “That is the essence of differentiation.”
