The most important question that companies face in deploying Artificial Intelligence (AI) is not technological but organizational: Should they use AI to increase the power of high-up managers or liberate front-line workers?
I suspect that the bulk of them will give the wrong answer to the question — and that we will be dealing with the consequences of their mistakes for decades to come, not just economically, as companies lose their creative flair, but also politically, as professional elites join the ranks of the angry and alienated.
Companies will evolve in radically different directions according to the answer that they give to this question. Choose the first answer and managers will use AI’s growing powers to divide jobs into identical units, monitor and measure workers in terms of their ability to fulfill their assigned roles and get rid of surplus workers. The faster you work, the more you will be rewarded.
Two directions
In the early 20th century the father of scientific management, Frederick Taylor, taught employers the importance of standardization and measurement of workflows. But today’s digital Taylorism ensures that the managerial eye is all-seeing, enabling employers to monitor not just your every movement but your fleeting emotions. And it hands enormous power to algorithms that are untouched by human emotion.
Choose the second, however, and workers will be able to use it to do remarkable things. They will improve the quality of their jobs by automating routine tasks (organizing their diaries or booking travel) but also improve the quality of their organizations by collaborating with other employees.
The first approach is superficially much more appealing because it provides companies with lots of low-hanging fruit. You can get rid of surplus employees: This week, Amazon, one of the most enthusiastic practitioners of digital Taylorism, announced that it is cutting its corporate workforce by 14,000.
You can satisfy your worry that too many workers are skiving by monitoring their comings and goings. You can reward “performance” rather than “presence” by measuring workers’ precise contribution.
Yet eventually the approach will fail. Studies of human motivation are remarkably consistent about what people want from organizations: They want to feel that they belong to a community, that their contribution is valued, that they can grow in their jobs, that they have a chance to exercise their creative faculties and that they are not constantly being micro-managed by higher-ups. Most people are willing to sacrifice a certain amount of pay if they can get these things.
Management gurus predicted that the digital age would lead to the triumph of a new age of entrepreneurial capitalism. In the new edition of “Humanocracy,” Gary Hamel and Michele Zanini demonstrate that we got the opposite — the triumph of top-down managerialism.
The number of people classified as “managers” or “administrators” doubled from 1983 to 2024, even as the rest of the workforce increased by 40%. The proportion of U.S. employees working for companies with more than 5,000 employees increased from 28.8% in 1987 to 36.4% in 2021.
The result of growing bureaucratization: 51% of U.S. employees claim that they are not engaged with their work while 16% claim that they are “actively disengaged.”
Dispiriting AI
So far, this pattern is being repeated with AI. Amazon, for example, has been dogged by accusations of insensitive management despite paying higher than average wages.
A 2020 study found that Amazon warehouse workers suffered from almost twice the rate of serious injuries as the industry average. But even more dispiriting is the company’s monitoring of every tiny detail of your daily routine, including the amount of time that you spend in the bathroom.
Similar complaints follow wherever the algorithms are unleashed. Truck drivers complain that their employers use AI-enabled tracking tools to spy on them. Fast-food employees complain that their schedules are being determined by algorithms, which measure demand, rather than by conventions such as regular hours.
Perhaps there is a limit to how much autonomy you can enjoy in providing taxi services or stacking boxes. But management by algorithm is being introduced into professional services that have traditionally placed a premium on self-management and self-improvement.
The same pattern is being repeated. Companies are exerting ever more minute control over employees: They can measure your keystrokes per minute, figure out how “collegial” you are in Zoom calls and bundle dozens of different performance measurements to determine your pay.
Companies are also extending automation into intimate human judgments. JPMorgan Chase & Co., which is the pacemaker in using AI in the banking sector, with an AI budget of $2 billion a year, allows managers to use AI to write performance reviews.
Sparks of resistance
There are some sparks of resistance. Young workers can vote with their feet by choosing companies that use AI to empower them rather than turn them into cogs in a machine. Some companies are recognizing the dangers of, say, automating entry-level jobs that provide you with your future stars.
Yet the race to adopt AI is so fast-paced and the FOMO so all-consuming that companies, in their obsession with hitting the metrics, are forgetting about subtle things such as creativity. And AI-driven management is self-reinforcing: The more the smart machines take over, the more skills atrophy and self-direction withers, creating yet more demand for smart machines.
The managerial eye is becoming at once ever more penetrating and ever less discerning.
Adrian Wooldridge is the global business columnist for Bloomberg Opinion and author of “The Aristocracy of Talent: How Meritocracy Made the Modern World.” His previous article was “The 19th century railroads show us the future of AI.”





Comments
Welcome to Post-Gazette subscriber-only commenting community. Please review our community guidelines.