TL;DR
Recent reports suggest that many tech CEOs are experiencing ‘AI psychosis,’ overestimating AI’s current and near-future capabilities. This phenomenon is linked to aggressive AI-driven layoffs and strategic decisions that may not align with actual AI performance.
Multiple industry sources indicate that several tech CEOs are exhibiting behaviors consistent with what some experts are calling ‘AI psychosis,’ overestimating AI’s capabilities and making strategic decisions based on these inflated perceptions. This phenomenon is raising concerns about organizational stability and the accuracy of AI-driven business strategies.
According to Aaron Levie, founder of Box, many CEOs are prone to ‘AI psychosis’ because they lack direct experience with the detailed work AI systems perform. Levie, who actively advocates for AI, argues that CEOs often see only the optimistic outcomes of AI prototypes, ignoring the complexities and limitations that frontline workers face. This disconnect leads to overconfidence in AI’s potential, prompting drastic measures such as mass layoffs and ambitious productivity claims.
Recent industry data shows that in 2026 alone, nearly as many layoffs have occurred as in all of 2025, with AI cited as a primary reason by many companies. Notably, Zeb Evans, CEO of ClickUp, announced a 22% reduction in staff after deploying around 3,000 AI agents, claiming the goal was not cost-cutting but creating a ‘100x organization.’ However, research from UC Berkeley, MIT, and Harvard suggests AI productivity gains are often overstated, with many studies indicating no clear, consistent link between AI adoption and increased efficiency.
Experts warn that AI models are still far from replacing human judgment and quality work across many tasks. Current projections estimate that AI systems will only reach an 80-95% success rate on text-related tasks by 2029, with full outperformance of humans taking several more years. Meanwhile, research indicates that the bottleneck in productivity has shifted to decision-making and approval processes, which are predominantly managed by executives, many of whom may overestimate AI’s readiness to handle organizational responsibilities.
Why It Matters
This trend matters because overconfidence in AI could lead to misguided strategic decisions, destabilizing organizations and wasting resources. The phenomenon of ‘AI psychosis’ highlights the risk of leaders acting on incomplete understanding, potentially causing chaos in organizational structures and workflows. As AI continues to evolve, ensuring that leadership maintains realistic expectations is crucial for sustainable growth and innovation.
The AI-Powered Professional: AI Productivity for Business Professionals Without the Technical Overwhelm (The AI-Powered Professional Series Book 1)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
In early 2026, the tech industry saw a surge in layoffs, with nearly as many cuts as the previous year, many linked to AI deployment. Prominent CEOs, like Zeb Evans, publicly credited AI for these decisions, often emphasizing productivity gains. However, academic studies from UC Berkeley, MIT, and Harvard have shown that current AI models do not consistently improve productivity and often fall short of expectations. Industry commentary, including Levie’s, suggests that many leaders are overestimating AI’s capabilities due to limited direct experience with the technology’s practical limitations.
“CEOs are uniquely prone to AI psychosis because they’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI.”
— Aaron Levie
“We rolled out about 3,000 AI agents to do internal work, and I laid off almost a quarter of our staff. This isn’t about costs; it’s about creating a 100x org.”
— Zeb Evans
“Studies indicate no robust relationship between AI adoption and productivity gains; models will need several more years to outperform humans reliably.”
— Research, UC Berkeley, MIT, Harvard
What Remains Unclear
It remains unclear how widespread the phenomenon of ‘AI psychosis’ is among tech leadership, and whether these behaviors will lead to long-term organizational instability. The pace of AI development and the accuracy of leaders’ perceptions are still evolving, making future impacts uncertain.
What’s Next
Industry analysts expect ongoing scrutiny of AI’s real-world capabilities, with potential regulatory and organizational responses to prevent overconfidence. Further research and case studies will clarify the actual impact of AI-driven strategies and whether leaders will adjust their perceptions accordingly.
Key Questions
What is ‘AI psychosis’ among tech CEOs?
‘AI psychosis’ refers to the phenomenon where tech leaders overestimate AI’s current and near-future capabilities, leading to strategic decisions based on inflated perceptions of AI’s potential.
Are the layoffs in the tech industry primarily caused by AI?
Many companies cite AI as a reason for layoffs, but industry research suggests that actual productivity gains from AI are limited, and some layoffs may be driven by other factors or inflated claims.
How reliable are current AI models in replacing human work?
Research indicates that AI models are approaching 80-95% success on text tasks by 2029, but full replacement of human judgment and quality work will likely take several more years.
What risks does ‘AI psychosis’ pose to organizations?
Overconfidence in AI can lead to poor strategic decisions, organizational chaos, and resource misallocation if leaders act on inflated perceptions rather than actual capabilities.
Source: Hacker News
