Critical Thinking in the AI Age

Sarkar opens with an AI-assisted workday: summarised email, generated replies, drafted reports, automated analysis and instant prototypes. Every task gets completed, yet the worker spends little time with the material. He calls this pattern “outsourced reason”: the person increasingly validates machine output and spends less time forming a view through reading, writing and analysis.

He surveys four possible costs. AI-assisted groups can converge on a narrower range of ideas. Workers report less critical-thinking effort when they trust AI more than their own judgement. People may remember less when a system performs the reading or writing. The workflow also creates a demanding management layer because the user must define the goal, divide the task and verify the result.

His alternative is an AI tool for thought. The prototype keeps the user inside the source material, supports strategic close reading, offers critiques and counterarguments, and lets the user build an argument outline. Generated prose follows those substantive decisions.

The practical principles are to preserve direct engagement, introduce useful resistance and prompt reflection on reasoning. A classroom application could ask students to compare sources, state assumptions, build an argument, answer counterexamples and explain which AI suggestions they rejected.