Blasé Capital HUMAN AI

If a recent survey is to be believed, it is humans who may derail the growing use of Artificial Intelligence (AI). A software firm concludes that while senior management and executives love it, the mid-managers and lower employees shun it, or pay lip-service to it. Almost 90 per cent of the first section use AI at workplaces, and senior management sings praises in earnings calls. In recent times, according to an article in ‘The Economist,’ “two-thirds of executives at S&P 500 companies mentioned AI. At the same time, the people actually responsible for implementing AI may not be as forward-thinking, perhaps because they are worried about tech putting them out of a job.” So, less than 60 per cent of the mid-managers, and less than 30 per cent of the employees use AI on the job. This implies that while these two categories happily go ahead with the AI-related efforts to please their seniors, they quietly refuse to use the tools.

A similar thing happened with desktops, laptops, and mobiles. In India, for instance, desktops inundated the state-owned banks and finance firms, as they did in progressive and forward-thinking private firms. Employees used them mostly to play games in the late 1980s, and early 1990s. In some cases, especially at the banks, desktops were purchased because of the top management’s decisions, but remained wrapped in original packaging for months, and years. The same happened with the laptops, and mobiles, when they became office tools to work, and communicate. Only over time, when there was no escape, and the Internet and web became a necessity did employees and mid-managers upgrade skills to handle the new tools, and new ways of working. Hence, AI may go through the same learning tech curve before the bulk of the employees adopt it actively, and adapt to its social and job disruptions at workplaces.

Yet another piece of insight lies in perception and imagination, which is subjective. New technology, whether it is hardware, Internet, or AI initially may lead to extra work, lower productivity, and lesser efficiency. In the case of desktops, and laptops the employees needed to learn new skills, which were more easily acquired by the youngsters. In the case of mobile-as-moving-offices, it was excruciating to work with the small screens to read, write, and monitor work. AI leads to such delays. Rewiring offices, as was the case with desktops, and making them wireless (Internet and mobile) takes time. Processes and workflows need to be re-defined. Employees need to get used to the newer ways. This may take months, or years. Erik Brynjolfsson of Stanford University dubs it the ‘J-curve’ of technology. Yvonne Chen, and others of ShanghaiTech University refer to it as “gen-AI’s mediocrity gap.” AI helps the weak workers, and does the opposite with the better ones, who may decide to work less since they can get the same work done faster.

When it comes to small and mid-sized firms, tech-led gains in productivity and profitability do not come naturally and logically. A recent poll by a global think tank finds that 45 per cent said that the returns from the use of AI were below their expectations. Only 10 per cent agreed that their expectations were exceeded. The same goes for profits, as most entities conclude that tech has not “significantly affected enterprise-wide profits.” This may be due to a dual strain of human nature. In most cases, we overestimate returns, and thus find that the benefits are not adequate. In addition, since we need to rewire the workflows, the transition takes more time than expected, and we get frustrated with tech. The small things matter, which change our subjective perceptions. For example, in the early period of desktops in India, it took a minute or two for the machines to start. This was deemed to be a waste of time compared to the old days when one could start work as soon as one sat at the desk.

For the experts, this translates into either non-adoption of AI, or delays in adoption. Some surveys hint that this is indeed the case in the US. “John Hartley of Stanford University and colleagues found that in September (2025) 37 per cent of Americans used generative AI at work, down from 46 per cent in June…. Ramp, a fintech firm, finds that in early 2025 AI use soared at American firms to 40 per cent, before levelling off. The growth in adoption really does seem to be slowing,” states ‘The Economist’ article. But there can be human-related reasons for this slowdown. When new tech arrives, there is the initial excitement to try, and experiment with it. Top management forces managers, and employees to try it out. Within a few months, many of the users get frustrated, and the number of new people opting for it begins to wane. This may explain why the adoption percentages have come down steeply within months.

Even if the adoption is accelerated due to internal and external pressures, ‘The Economist’ piece gravely concludes that the “pause suggests that the economic pay-off from AI will arrive more slowly, more unevenly, and at a greater cost than implied by the current investment boom.” Yet, until adoption accelerates, the “revenues required to justify $5 trillion in AI capex will remain out of reach.”