AI automation reduces employment and labor income
Eroding economic independence deepens dependence on technology owners
Concentrated wealth raises concerns about widening disparities in political influence

Warnings have emerged that the spread of autonomous weapons and artificial intelligence (AI) surveillance networks could concentrate power and lead to “digital authoritarianism.” The concern is that democratic checks on power will weaken if political leaders or major technology companies monopolize the means to monitor citizens and suppress resistance. These concerns are gaining a factual basis as technology increasingly replaces human labor and military personnel. Companies continue to cut jobs in the name of AI adoption, while unmanned equipment has begun taking over soldiers’ duties on the battlefield. As governments and companies become less dependent on citizens’ labor and military service, they are also likely to face less pressure to accommodate citizens’ demands.

Daniel Suarez: “AI-Driven Concentration of Power Threatens Citizen Oversight”

According to Nikkei Asia on the 26th, American science fiction author Daniel Suarez warned that the spread of autonomous weapons and AI surveillance networks risks bringing about digital authoritarianism. His focus was on a situation in which those in power monopolize the means of surveillance and lethal force and escape citizens’ control. The concern is that democratic checks will weaken if those in power can continuously track citizens’ movements and secure the means to suppress resistance by force. Suarez cited an extreme version of this future in which rulers declare that they no longer need elections or democracy.

The actors behind the concentration of power he cautions against include both political leaders and major technology companies. He expressed concern about a system in which a small group controlling AI surveillance infrastructure and autonomous weapons can wield enormous coercive power without being accountable to the public. Surveillance networks risk being abused to identify opponents of those in power, while killer robots could be used to suppress resistance. His warning is that combining the two technologies would make it harder for citizens to challenge abuses of power or hold those responsible accountable.

Suarez emphasized the rule of law and human intervention as principles for preventing this outcome. He explained that what matters is who receives the authority to decide on surveillance and the use of force, and what rules constrain that authority. In particular, he advocated a “human-in-the-loop” principle, under which people remain involved in autonomous weapons’ decision-making processes. Warning against ceding authority to AI while consumed by technological competition, he called for both legal controls on those operating the technology and the involvement of human judgment.

AI Replacement Pressure Hits Both Youth Hiring and Existing Workforces

The future Suarez warned about is not an impossible scenario. If AI replaces human intellectual labor, companies that own the technology could reduce their dependence on staff, while workers could struggle to retain their jobs and incomes. This would upset the balance of power between those who own technology and those who supply labor. In an interview with Semafor in August, Microsoft (MS) co-founder Bill Gates predicted that AI would replace tasks across multiple industries simultaneously, sharply reducing the number of jobs available. If even the occupations workers might move into are automated, as he feared, their employment options will narrow.

An early sign in the labor market has been a decline in employment opportunities for young people entering the workforce. An analysis of payroll provider ADP’s data by Stanford University’s Digital Economy Lab found that, as of June this year, employment among 22- to 25-year-olds in occupations highly exposed to AI was about 19% below the level expected had it followed the same trend as occupations with low exposure. The researchers attributed the widening gap to companies cutting back on new hires of young workers. Employment declines were concentrated in occupations where AI is used to automate human tasks, with young workers who rely on standardized knowledge learned from textbooks or documents particularly affected.

As entry barriers rise for young job seekers, companies are also moving to reduce their existing workforces in the name of AI adoption. After cutting about 30,000 jobs between late last year and January this year, Amazon reportedly eliminated fewer than 1,000 additional positions on the 7th of this month, mainly in its Stores division, which handles e-commerce, and other units. Meta is pursuing a restructuring that would cut about 8,000 jobs this year and has also eliminated 6,000 planned positions. It proposed reassigning 7,000 employees to AI-related organizations and reducing management roles. Microsoft also announced in July that it would cut 4,800 jobs, about 2.1% of its total workforce, mainly in its commercial business, Xbox division and other units. Oracle’s total headcount fell by about 21,000 during fiscal 2026, which ended in late May this year, and it began additional layoffs in September. Netflix announced a restructuring on the 14th of this month that would eliminate 1,000 jobs, about 6% of its workforce, including positions in product, technology, legal and human resources.

Table 1. Layoffs and Organizational Restructuring at Major Technology Companies

Ground Robots Take Soldiers’ Places, From Supply Runs to Combat

The trend of physical AI replacing tasks performed by people is also emerging in the military. Ukrainian forces are using ground robots to transport supplies and evacuate wounded soldiers, carrying out more than 9,000 such missions in March alone this year. At the time, Ukraine’s Defense Ministry pursued contracts to purchase 25,000 unmanned ground vehicles in the first half of the year, with the goal of handling all front-line military logistics using robots. The aim is to reduce how often soldiers must enter areas where they are exposed to enemy attacks. An Independence Day military parade held in Kyiv on August 24 featured ground robots for combat, reconnaissance, supply and evacuation missions, along with aerial drones and unmanned surface vessels. Dozens of ground robots paraded through the city center while unmanned vessels maneuvered on the Dnipro River, showcasing equipment that has taken over soldiers’ duties.

Ground robots are also being deployed on combat missions to repel enemy attacks and destroy military facilities. The NC13 unit of Ukraine’s 3rd Army Corps, which Reuters visited in September, conducted combined operations using machine-gun-equipped unmanned ground vehicles to suppress Russian attacks while striking the enemy from the air with one-way attack drones. Soldiers operate the machine guns remotely from a command post while watching battlefield footage transmitted by the robots. In one instance, a ground robot towed a trailer carrying 3 bombs to blow up a bridge used to supply Russian forces. The unit’s deputy commander said robots were being deployed on operations almost daily and that at least 2 Russian soldiers had surrendered to ground robots. The 3rd Army Corps set a goal of replacing about one-third of its front-line troops with unmanned ground vehicles by the end of the year.

Generative AI Also Used to Build Tactical Map Analysis Software

As Ukraine deploys robots on the battlefield, a case has also emerged in the United States of generative AI being used to build military software. In an example disclosed in June by the Massachusetts Institute of Technology (MIT) Lincoln Laboratory, U.S. Air Force cadet Joshua Lynch, who had no coding experience, developed a military software prototype with AI’s help. Someone who had identified functions needed in military operations attempted to implement them directly without formal programming training. Under the guidance of laboratory researcher Laura Niss, Lynch described the functions he needed in everyday language to ChatGPT, Claude and Gemini, asking them to write and revise the code. Built over about 3 months using this so-called “vibe coding” approach, ROMAD-AI included functions for analyzing tactical maps and generating mission-planning documents.

During development, Lynch focused on communicating detailed requirements to AI and checking its output. When the AI lost track of task priorities or modified unrelated code, he broke the problems down into smaller pieces and asked questions about each. He reduced errors by steering conversations back on course whenever they drifted away from the development goals. Much of the project was also spent identifying AI’s limitations and learning how to obtain the desired results. Niss highlighted the significance of a service member with no coding knowledge turning operational needs into a working program. She assessed it as a useful tool for people familiar with military work to communicate required functions and solutions in concrete terms to technical experts.

Work, Taxation, Defense and Education on Shaky Ground

As AI takes on even development work requiring specialized knowledge, uncertainty is growing around the path by which people obtain an education, find jobs and earn a living. If demand for workers in a given role declines before they can recoup the time and money spent on education, the route to a better standard of living through academic qualifications and skills narrows. Even if automation boosts productivity, there is no guarantee that the gains will translate into higher worker incomes. In a study analyzing AI’s economic effects, MIT professor Daron Acemoglu projected a widening gap between capital income and labor income. As companies cut staff to reduce production costs, some of the wages previously paid to workers shift into returns for capital owners.

Declining labor income also risks weakening governments’ fiscal capacity to support the unemployed. As more people lose jobs or suffer wage cuts, income tax and social insurance contribution revenues also come under pressure. International Monetary Fund (IMF) researchers have pointed out that the spread of labor-replacing automation could weaken the tax base just as spending needs for unemployment support and retraining increase. Their analysis suggests that, particularly in tax systems where effective rates on capital income are lower than those on labor income, the shift of income toward capital owners makes it difficult to fully offset lost tax revenue. If fiscal pressure prevents adequate retraining and livelihood support, unemployed people must shoulder both the cost of acquiring new skills and living expenses during their job search. This raises concerns about a vicious cycle in which an inability to afford training delays reemployment, while longer periods without income make returning to the labor market even harder.

If the erosion of economic independence is compounded by the shift to unmanned military systems, citizens’ influence over governments and companies could also diminish. While workers participated in production and citizens performed military service, securing their cooperation was a major concern for business operations and national security. But as production facilities and military operations become increasingly unmanned, work stoppages and refusals to mobilize are also likely to lose their effectiveness as means of pressuring those in power. The worst-case scenario outlined by historian Yuval Harari similarly envisions a situation in which the majority displaced by AI lose both their economic roles and their political influence. His concern is a future in which people excluded from employment become dependent on a small group that owns wealth and technology, while those in power suppress resistance through constant surveillance.