AI Is Making Us Dumber – Unless We Fight Back

The piece argues that generative AI, while delivering real efficiency gains, risks eroding critical thinking through "cognitive offloading" when people accept machine answers without question. Empirical research led by Monideepa Tarafdar with 45 knowledge workers at a U.S. public university shows a practical alternative: use AI to generate friction by asking it to contradict one’s assumptions, play devil’s advocate, or surface unlikely perspectives. Workers who deliberately prompted models to challenge their views performed significantly better; earlier human-AI collaboration studies in loan evaluation reported similar gains. The core claim is that AI’s value lies less in turnkey answers than in forcing humans to ask harder questions and exercise judgment.

Concrete examples illustrate the point: a marketer used AI-generated negative customer personas to reveal product improvements (e.g., using cross-industry case studies), an attorney used AI to surface obscure loopholes and realistic misuse scenarios, and an operations researcher adapted methods from other fields suggested by AI to strengthen models. Creative writers produced stronger copy when treating AI as a sounding board rather than a ghostwriter. Practical recommendations for organizations include teaching friction-generating prompts, running AI bootcamps and sandboxes, fostering cross-functional sharing, and appointing AI champions. The recommendation is clear: treat AI as an intellectual sparring partner that amplifies human judgment, not a replacement, to avoid societal atrophy of reasoning skills.