A survey of nearly 6,000 business executives in the US, UK, Germany and Australia found that more than 80% saw no measurable impact of AI on employment or productivity.
The National Bureau of Economic Research (NBER) study in Massachusetts is based on input from CFOs, CEOs, and executives from companies of various sizes in four countries.
On average, 69% of companies currently use some form of AI, and 75% expect to use AI in the next three years. Typical uses include “text generation using large-scale language models,” followed by “visual content creation” and “data processing using machine learning.”
According to those surveyed, AI will have limited impact on employment and productivity. More than 90% of managers say AI has had no impact on employment in their organization over the past three years, and 89% say they have seen no change in productivity (as measured by revenue per employee).
Nevertheless, many executives expect significant impacts, including job losses, over the next three years. NBER estimates that this will hurt about 1.75 million jobs in all four countries by 2028.
Respondents also expect AI to increase business productivity by approximately 1.4% over the next three years. This suggests a reversal of the long-term decline in productivity growth in many developed countries, the authors said.
The report also notes that employees’ expectations differ from those of senior executives, with workers expecting more jobs to be created as a result of AI over the next three years but only modest productivity gains.
The NBER paper adds to a growing body of evidence that the commercial benefits of AI adoption are not living up to their promise, at least not yet. A recent survey of more than 4,500 business leaders conducted by PwC consultants found that more than half reported not seeing revenue increases or costs decrease.
A study by professional services firm Deloitte found that 74% of organizations want to increase revenue through AI initiatives, but only 20% actually do so.
A trial of Microsoft’s M365 Copilot by a UK government department announced in September found no productivity gains, speeding up some tasks while hindering others.
At the same time, Jared Spataro, the executive leading Microsoft’s AI at Work effort, acknowledged that he struggles to emphasize Copilot’s return on investment (ROI) because much of the knowledge work doesn’t directly translate into sales or bottom-line numbers.
But that hasn’t stopped others within Microsoft from making bold claims. Just this month, Microsoft’s head of AI, Mustafa Suleiman, said that most tasks that involve “sitting down at a computer” will be fully automated by AI within the next year or year and a half. He cited accounting, legal affairs, marketing and project management, according to Fortune magazine.
Lenovo recently claimed that companies in Europe and the Middle East are accelerating the adoption of AI, with 94% of companies surveyed expecting to see a positive return on their investments, despite growing evidence to the contrary.
A study released today by Gartner found that organizations expect AI to transform customer service and improve the customer experience as humans continue to make decisions.
Additionally, 91% of customer service leaders say they are under pressure from management to implement AI.
Nearly 80% of companies plan to place at least some agents into new roles as routine tasks are expected to be automated, but human expertise will still be needed for “complex or emotionally sensitive” interactions. According to Gartner, 84% of leaders plan to add new skills to agent roles to support this change.
All of this points to the possibility that the productivity gains from AI adoption will be negligible, in stark contrast to the hundreds of billions of dollars that tech giants have poured into developing these systems. With traffic expected tomorrow, it’s no wonder they’re all keen to endure the rollout. ®
