The technology landscape is increasingly dominated by artificial intelligence (AI), with major tech corporations releasing sophisticated digital assistants almost daily. Platforms like Meta‘s Muse, Google‘s Gemini, and OpenAI’s dots promise to streamline life and work by automating processes and handling tedious tasks. While these tools initially appear helpful and harmless, some researchers warn that this convenience paves the way for a deeper, more subtle threat to human society.
Experts caution that the danger extends beyond simply job displacement. They warn that the erosion of human skills and the surrender of critical thinking to AI poses an existential risk to humanity. This danger is termed “gradual disempowerment.”
Understanding Gradual Disempowerment
Gradual disempowerment describes a trajectory where AI systematically takes over human labor in the economy, assumes control of key societal decision-making processes, diminishes human creativity, and degrades human cognitive abilities over time. According to experts, the word “gradual” should not be interpreted as meaning the problem is distant.
Jay Chooi, founder and CEO of Robocurve, a company that monitors AI-driven robotics, stated that he places “gradual disempowerment” near the top of his concerns. He noted that this scenario is as worrisome as the potential for engineered pandemics or advanced cyberattacks, which could be accelerated by capable AI.
“It’s a civilization-scale problem,” Chooi said. “A lot of people feel very powerless, because it requires a lot of effort and a whole-of-society approach to solve it.”
How AI Facilitates Disempowerment
The concept of gradual disempowerment originated from a paper published in early 2025. The core concern is the rapid advancement of AI capabilities, which have accelerated since the launch of ChatGPT in 2022. AI models are rapidly becoming superior to humans in nearly all cognitive areas—including complex mathematics, hard sciences, coding, writing, and artistic creation.
Furthermore, advances in robotics mean that AI-powered machines are outperforming humans in physical labor, such as construction, mining, manufacturing, and even detailed tasks like painting murals. This creates a scenario where AI agents or robots are superior to humans in every aspect of a job—moving faster, thinking more deeply, and requiring less maintenance than human staff.
To illustrate this economic pressure, consider a fictional manufacturing company. If a competitor adopts AI-powered machines that can build goods much faster and cheaper than human workers, the original company faces an impossible choice: fire its staff or face bankruptcy. This free-market incentive structure pushes companies toward automation, accelerating the process of disempowerment.
Impacts on Human Cognition and Governance
As humans cede more tasks to AI, they become less skilled at performing those tasks themselves. Margaret Mitchell, an AI research scientist at Hugging Face and co-author of a recent paper on retaining human involvement in AI, warned about this dependency. “If people aren’t actually practicing the skills that the agents are doing for them, they lose the ability to have that skill, and so now you become reliant,” she stated. She questioned, “What happens when something goes wrong? There’s no human you can point to who can fix it. So now you’ve become fully dependent on these systems to fix themselves, and what if that’s not possible?”
The risk extends beyond the economy into governance and the military. Currently, human decision-making is central to warfare, though AI contributes to the process. If nations rely heavily on AI for military decisions, those less reliant on technology will be at a disadvantage. Jan Kulveit, a co-author of the disempowerment paper and co-founder of ACS Research Group, noted the critical role of citizenship. “The states need citizens. They need us because we pay taxes and we fight for the states in the military,” he said.
Kulveit suggested that if governments no longer require citizens for labor or military defense, they become less likely to function democratically. He compared theoretical AI-run states to current oil states, arguing that if a state doesn’t depend on its populace for funding or support, it tends toward authoritarianism.
Strategies to Counter Disempowerment
The authors of the gradual disempowerment paper argue that this transition is not necessarily the result of a deliberate scheme by AI, but rather an economic inevitability driven by the perceived value AI adds to commerce and state functions. They argue that humanity cannot survive by merely competing with machines.
To mitigate this risk, several strategies are proposed. Firstly, it is vital to track the degree and type of job displacement caused by AI, requiring detailed indicators of which specific roles are being automated.
Secondly, proactive measures must be taken to keep humans involved in the AI process. Margaret Mitchell argued that the industry tendency to seek “frictionless” systems is detrimental. Instead, she advocated for “strategic friction”—designing AI systems to require human input at key checkpoints. This approach can draw attention to critical issues that might otherwise be overlooked.
Kulveit echoed this, suggesting that AI should be designed to be complementary to human capability rather than purely substitutive. While quick answers are useful, they can foster over-reliance. He suggested that AI should function not just by providing answers, but by educating the user, thereby allowing the human worker to learn alongside the technology and remain a partner, not a mere subordinate.
Despite the relative lack of dramatic urgency compared to threats like bioweaponry, experts maintain that gradual disempowerment is a highly visible and imminent risk. Jay Chooi stated that “All the organizations have the incentives to adopt it and phase out humans,” calling it a very logical conclusion.”
