A new UCLA research study added an unexpected dimension to the AI adoption conversation this week: what happens to people when you take their AI tools away? The findings suggest the dependency may be more significant than most organizations have accounted for.
UCLA researchers recruited about 1,220 people, gave half of them AI assistants for reasoning tasks, and then yanked the AI away mid-experiment. What happened next is the behavioral equivalent of putting someone in a wheelchair for a month and then being shocked when their legs feel wobbly. Neuralbuddies
Participants who had AI assistance showed measurable declines in independent reasoning performance once the tools were removed - a finding the researchers describe as cognitive offloading with dependency effects.
Why This Matters for Enterprise AI Strategy
The study surfaces a risk that most enterprise AI rollouts have not formally addressed: workflow dependency. When an organization integrates AI deeply into daily work, the productivity gains are real - but so is the exposure if the tool becomes unavailable, the vendor changes terms, or the organization needs to switch platforms.
This is not an argument against AI adoption. The productivity benefits of AI tools are too significant to pass up based on hypothetical disruption scenarios. But it is an argument for building AI literacy alongside AI adoption - ensuring your team understands what the AI is doing and can operate effectively without it when necessary.
The Broader Context
Stanford's 2026 AI Index cites Pew data showing 56% of AI experts believe AI will have a positive US impact, versus just 10% of Americans feeling more excited than concerned about AI. Neuralbuddies
That gap between expert optimism and public anxiety is, in part, a reflection of these dependency concerns playing out at scale. The public is not wrong to think carefully about what it means to rely on tools that can be changed, restricted, or taken away.
What This Means for Your Business
From my time advising organizations on AI implementation, the best AI deployments share one characteristic: they treat AI as an amplifier of human capability, not a replacement for it. The UCLA findings reinforce what the strongest enterprise AI teams already know - you need AI fluency, not just AI access. Train your people to understand the tools, not just use them. That's what separates the leading cohort from the rest.
