Featured image: A man is sitting at a wooden desk. AI-generated image; not a photograph of the reported event.
Knowledge workers—such as scientists, teachers, lawyers, and business leaders—are increasingly using generative AI for creative and problem-solving tasks.
Cognitive Offloading: A Risk in AI Use
Users may surrender to AI’s quick, confident answers.
This is known as cognitive offloading.
It occurs when people rely on AI to do mental work instead of thinking through problems themselves.
Such reliance can weaken critical thinking—skills that make knowledge workers valuable.
Friction-Driven AI Use Improves Outcomes
Research from UMass Amherst shows a different result: when users challenge their ideas with AI, performance improves.
Researchers interviewed 45 professionals at a U.S. public university who used AI in core tasks like idea generation and problem-solving.
They analyzed AI prompts and outputs from iterative chats and found that opposing viewpoints led to stronger, more creative results.
How to Create ‘Friction’ with AI
- Ask AI to generate scenarios that contradict your assumptions.
- Request examples from different industries or contexts.
- Challenge AI to find loopholes or edge cases in your logic.
A marketing professional asked AI to create virtual customers who would dislike an accounting certificate program.
The feedback revealed preferences the designer had not considered—like using cases from multiple industries, not just one.
An attorney used AI to find legal loopholes. That exposed realistic, hard-to-detect unethical behaviors in employees.
AI as a Tool for Deeper Understanding
AI can help users examine concepts from multiple angles.
A research scholar used AI to find journal papers applying a concept in different settings. That deepened his understanding.
A lawyer followed up on AI-suggested citations by asking for varied case examples. That helped her fully grasp the original case and write a strong brief.
An operations research scientist adapted unfamiliar programs—like those from signal processing—suggested by AI. His strong background in supply chains helped him apply them effectively.
Why This Matters for Work and Learning
AI is now central to many professional workflows.
Its value depends on how it is used.
When used to challenge assumptions, AI becomes a thinking partner—not a shortcut.
How Organizations Can Support This Practice
Leaders can promote questioning prompts over agreeing ones.
- Host internal workshops on AI use and limits.
- Offer AI sandboxes—safe, private spaces for experimentation.
- Run regular sessions led by cross-functional AI task forces or champions.
- Create competitions with prizes to boost engagement.
A professor in AI education led a program with prizes for professors who completed AI bootcamps.
Key Takeaway: Friction Helps AI Work for You
Adding friction may seem like making AI work against you.
But research shows it actually improves AI’s effectiveness—by revealing blind spots and sparking deeper analysis.
This article is republished from The Conversation under a Creative Commons license.
Monideepa Tarafdar, Ph.D., is professor of information systems at UMass Amherst.
Explore more on AI in work: AI in the workplace and AI and job skills.
Learn how AI can support human thinking in education: Community programs and social outcomes.
