Why companies might stop employing people
For employers, the reasoning will be straightforward: AI offers labor that is free or almost free, according to Yampolskiy. “If I can just get a $20 subscription or a free model, it makes no sense to hire humans for most jobs,” he said.
Companies may completely replace employees with humanoid robots to do physical labor and cognitive ones like data processing and coding. AI systems that operate continuously might generate value worth millions of dollars.
Additionally, physical occupations might not be safe
According to Yampolskiy, humanoid robots are just five years behind software in development, despite the widespread belief that manual work is immune to automation.
By 2030, physical labor including driving, warehouse jobs, and simple construction may be automated. This robotics wave, when paired with software-based AGI, has the potential to replace both white-collar and blue-collar jobs.
If this were to be the case, it would be the first time in history that almost every industry would no longer require human labor.
High unemployment rates are anticipated
Yampolskiy cautions that we are not discussing unemployment rates of 10% or even 25%. He estimates that just a small portion of jobs will remain after a startling 99% job loss.
It’s possible that the few remaining jobs—like personal caring or some creative roles—are ones where people prefer to engage with other people.
However, in most areas, AI will surpass humans. This projection is significantly more severe than mainstream forecasts, which normally anticipate gradual transitions and partial automation.
What jobs may humans still have?
According to the professor, certain occupations may exist solely because human like other humans in those roles, rather than because they are necessary.
Personal care, emotional support, and entertainment may continue to include individuals on their own initiative. However, even these may be challenged by more human-like robots and AI partners.
Teaching, therapy, and creative production may endure for cultural reasons, but Yampolskiy emphasizes that these are outliers in a society dominated by automated work.
Even prompt engineering and coding are vulnerable
Yampolskiy brushed off popular recommendations like coding or prompt engineering when questioned about “future-proof” careers. AI systems are already more adept than humans at creating prompts for other AIs, he claimed, and coding automation is developing quickly.
That means that even highly technical, freshly publicized occupations may not be secure. As AI evolves and improves, retraining may become ineffective. Instead, any ability may eventually be outrun, leaving workers with few options for security.
No retraining plan can keep up
According to Yampolskiy, this technical change will not enable effective retraining like earlier ones did. Some occupations were lost to automation in previous decades, but new ones were created.
Retraining is pointless though, if AI can perform any task. He said, “There is no plan B because all jobs will be automated.”
Because of the quick pace of AI advancement, the new area may already be out of date by the time people reskill. This weakens conventional policy responses to unemployment.
The professor’s extensive knowledge of AI safety
Given his extensive experience studying the challenges associated with AI, Yampolskiy’s projections seem credible. In 2011, he came up with the term “AI safety,” and he has written more than 100 papers.
The ethical dangers and computational frameworks for controlling superintelligent systems are examined in his 2025 book, Considerations on the AI Endgame.
Far from being a casual observer, he has dedicated his career to warning people about the perils of artificial intelligence. His grave prognosis is based on both scientific skill and years of cautionary study.
Comparing his opinions with those of experts
Other AI leaders anticipate significant impacts, despite Yampolskiy’s 99% estimate being exaggerated. White-collar jobs might be mostly automated, according to Geoffrey Hinton, who is sometimes referred to as the “Godfather of AI.”
In five years, up to 50% of entry-level office employment may disappear, according to Anthropic CEO Dario Amodei.
Even more upbeat voices, such as Sam Altman of OpenAI, recognize that there will be a lot of upheaval in the future. The primary point of contention is not whether AI will change professions, but rather how quickly and drastically.
Humanoid robots’ role in the transition
Although software automation may be the first to come, humanoid robots may shortly follow. Within five years, robots that can replace human labor will be ready, according to Yampolskiy.
According to technological developments from Chinese manufacturers, Boston Dynamics, and Tesla, mass manufacturing is almost at its peak.
When implemented, robots might do tasks including retail stocking, delivery driving, and even medical assistance. When AGI brains and robotic bodies are combined, human work may become economically obsolete in both the digital and physical realms.
Why the forecast might not entirely play out
Yampolskiy acknowledges that just because AI can automate almost everything does not imply it will happen overnight. Economic, social, and political constraints may impede adoption.
For example, firms may be hesitant to replace employees for public relations reasons, or governments may control deployment.
Historically, technologies that were theoretically feasible took decades to spread. However, his thesis is that the ability to replace people may exist far sooner than society is prepared to manage.
The risk of addicted idleness
Aside from the financial consequences, Yampolskiy argues that unemployment undermines people’s feeling of purpose. Jobs offer more than simply money; they also give structure, prestige, and community.
If most jobs disappear, communities must develop new methods to create purpose on a large scale. Otherwise, he warns, affluence may deteriorate into “addictive idleness,” with people trapped in entertainment loops rather than meaningful activities.
The social cost of widespread unemployment may parallel the economic one, changing how people find meaning in their everyday lives.
Potential substitutes for work in society
Yampolskiy proposes that if employment disappear, civilizations may need to develop artificial systems of meaning. Universal dividends generate income, civic service programs give structure, and cultural recognition systems replace employment-based status.
Local institutions and virtual communities may serve as alternatives to workplaces as sources of connection.
It will be vital to carefully design these systems. Without them, nations risk crumbling into inequity and despair since most people lose their livelihoods as well as the identities that come with them.
Governments may be unprepared for this transformation
A common element in Yampolskiy’s warnings is that governments are unprepared for a 99% jobless rate. Few countries have measures in place to deal with such rapid employment losses.
Safety nets, retraining programs, and labor regulations all presume partial automation, rather than near-total automation. Without proactive preparation, governments risk social discontent, economic collapse, and political instability. The professor’s warning will encourage governments to act before it is too late.
What this means for young workers
Yampolskiy’s projection is particularly worrying for young individuals entering the job market. Traditional recommendations to study in “safe” professions may no longer be valid if AI can outperform every talent.
Law, banking, and engineering careers, which were formerly thought to be secure, may be dwindling. Students may have to prepare not only for work, but also for a future in which jobs are scarce.
The unpredictability causes worry, but also emphasizes the significance of adaptation and lifelong learning.
See how Intel’s most recent layoffs underscore the changing environment of technology jobs and strategy.
