AI adoption slows in workplaces despite hype and massive investment

Corporate adoption of artificial intelligence (AI) appears to be slowing, raising questions about the pace at which the technology will deliver economic returns, according to a new analysis of data in The Economist. Figures quoted from the US Census Bureau suggest that the proportion of employees using AI at work has edged down to around 11 percent in recent weeks. The decline is most notable among larger organisations with more than 250 staff, where uptake had previously been stronger. The findings indicate that, three years into the current wave of generative AI development, business demand may be less robust than anticipated.

The rate at which companies adopt AI has significant implications. Much of the expected productivity gain from the technology depends on its integration into everyday business processes rather than limited or experimental use. It also underpins the commercial case for continued investment. Large technology firms are projected to spend around $5 trillion on AI infrastructure by 2030, with analysts suggesting this would require annual revenues of approximately $650 billion to be sustainable, compared to about $50 billion today.

Other studies present a mixed picture but point to a similar trend. Research from Stanford University found that the share of Americans using generative AI at work fell from 46 percent in June to 37 percent in September. Data from the Federal Reserve Bank of St Louis indicates only marginal growth in daily use, while separate figures from fintech firm Ramp suggest that adoption surged earlier in 2025 before levelling off.

Differences between surveys may reflect how AI use is defined and measured. Some economists argue that official data may understate adoption because of narrow definitions, while others suggest it provides a more representative view across industries, rather than focusing on sectors such as technology where uptake is typically higher.

Several factors may be contributing to the apparent slowdown, claims The Economist. Economic uncertainty, including shifting trade conditions and interest rate expectations, may be prompting organisations to delay investment. There is also historical precedent for uneven adoption of new technologies, with periods of slower growth followed by more rapid expansion.

Internal organisational dynamics may also play a role. While senior leaders frequently highlight the importance of AI, implementation often depends on middle management and operational teams, where enthusiasm may be more limited. Surveys indicate a gap between executive use of AI and that of managers and frontline employees, suggesting that adoption initiatives may not always translate into sustained practice.

There are also indications that expectations around the impact of AI are being reassessed. Market data shows that companies investing heavily in AI have not consistently outperformed their peers, while surveys of executives report that many AI initiatives have delivered lower-than-expected returns. Research from consultancies including Deloitte and McKinsey suggests that, for most organisations, AI has yet to make a significant contribution to overall profitability.

Academic studies point to additional challenges. Introducing AI can initially reduce productivity as organisations adapt systems and workflows, a phenomenon described as a “productivity J-curve”. There is also emerging evidence that the technology may standardise output, enabling average performance while potentially limiting higher levels of individual productivity.

Although AI capabilities are expected to improve and organisations are likely to refine their approaches, the current slowdown suggests that widespread adoption may take longer than anticipated. This, in turn, could delay the economic benefits associated with the technology and complicate efforts to justify the scale of ongoing investment.