Instead of asking AI to do your work, consider asking it to challenge your work.
Evidence suggests that heavy reliance on generative AI can reduce critical thinking and erode job skills. The problem of “cognitive offloading”—being too quick to outsource our mental workload to AI—can also lead to shoddy, subpar work.
But knowledge workers who use AI correctly can still gain a significant edge, writes University of Massachusetts (Amherst) Professor Monideepa Tarafdar in The Conversation. The key is to create “friction” by having AI challenge your ideas.
Researching opposing viewpoints has always been a good idea, but it matters even more today because AI models tailor their output based on users’ previous prompts and responses. This tendency can create an “echo chamber” effect that leads the models, and their users, to overlook information that doesn’t fit an established pattern.
To mitigate this vulnerability, Tarafdar says users should ask AI directly for opposing perspectives. For example, a marketing professional in her study asked AI to create virtual customers who disliked the product that the professional was marketing. The AI model’s negative feedback included new ideas for the product that its designers had not considered.
Another research subject, an attorney, asked an AI model to find obscure legal loopholes and game out how they might be exploited, which prompted it to provide “surprising but realistic” scenarios of unethical worker behavior.
Along with her own work, Tarafdar cites outside studies supporting the idea that creating “friction” with AI can improve its performance.
In a study on human-AI collaboration for loan evaluations, researchers found that human-AI collaboration produced better results than human-only or AI-only work, and that collaboration involving disagreement produced the best results.
Another study showed that creative writers produce better work when they use AI interactively as a sounding board, rather than to ghostwrite copy on its own.
Tarafdar says leaders should instruct their employees to use “friction-generating queries,” including by drawing on their own knowledge and experience to make the AI question its outputs.
“In these efforts to support knowledge workers, an important detail is pointing out that adding friction can appear to make an AI system work against you, but—as new research shows—it actually helps it work for you,” she writes.