The people who championed AI adoption across industries are now some of its most dependent users. A new study surveyed 200 UK business leaders and found that 62 percent use AI to make “most decisions,” according to Futurism AI, which flagged research conducted by market research agency 3Gem.
The irony is hard to miss. The executives pushing AI onto their workforces are quietly letting it do their thinking for them.
📊 The Numbers Tell a Clear Story
The 3Gem survey of business owners, founders, and CEOs revealed several striking data points:
- 62% use AI to make “most decisions”
- 70% (140 out of 200) second-guess their own ideas when they conflict with AI recommendations
- 46% now rely on AI advice more than input from their own colleagues
- 27% said they used AI for termination decisions in 2025
That last figure actually represents a decline from a similar 2024 report, where 64 percent of business leaders said they consulted AI on firing decisions. Still, the overall trend points in one direction: growing dependence.
🧠 The Science Behind the Brain Drain
This isn’t just anecdotal. A joint Carnegie Mellon and Microsoft study from last year found that knowledge workers who trusted generative AI’s accuracy showed lower propensity for critical thinking. The mechanism is straightforward: when humans believe a task has been competently automated, they disengage.
Danish psychiatrist Søren Dinesen Østergaard, who predicted what’s now called “AI psychosis,” warned in February that academic scholars risk building up “cognitive debt” by outsourcing their work to chatbots. The pattern holds across professions and seniority levels.
🔍 What This Means for Practitioners
For anyone working with or under AI-dependent leadership, a few practical takeaways stand out:
- Decision quality may be declining at the top. If your CEO is deferring to ChatGPT over their own team, the strategic direction of the company is only as good as the model’s output.
- AI should augment, not replace, judgment. The research consistently shows that treating AI as a thought partner (challenging its outputs, cross-referencing with human expertise) produces better outcomes than blind trust.
- Track your own dependency. If you notice yourself accepting AI suggestions without scrutiny, that’s the cognitive atrophy researchers are warning about. Build in deliberate friction: review AI outputs critically before acting on them.
- Teams should push back. When 46% of leaders trust AI over their colleagues, that’s a signal for teams to advocate more forcefully for human expertise and institutional knowledge.
⚠️ Limitations Worth Noting
The 3Gem study surveyed just 200 UK business leaders, which is a relatively small sample focused on one market. Self-reported data on AI usage also tends to be unreliable: respondents may overstate or understate how much they rely on these tools. The study was conducted by a market research agency, not an academic institution, so peer review wasn’t part of the process.
What stands out here isn’t any single finding. It’s the convergence. Multiple independent studies, from Carnegie Mellon, from clinical psychiatrists, and now from industry surveys, all point to the same conclusion: heavy AI reliance weakens critical thinking regardless of how senior or experienced the user is.
The executives who told everyone else to embrace AI are learning the same lesson the hard way. For more details, check the original reporting from Futurism AI.