[AI Ethics Seminar 2025 EP.1] How AI Is Changing Human Critical Thinking - LG AI Research Blog

While generative AI is maximizing productivity, concerns are being raised that as dependence on AI grows, humans may lose the opportunity to think for themselves. According to recent research, when using AI, the human role is shifting from a 'Doer' who directly defines and solves problems to a 'Steward' who reviews and corrects the output produced by AI.

In this process, the biggest risk is Over-reliance, uncritically accepting AI's output. To prevent this, Microsoft Research has proposed the concept of 'Appropriate Reliance (AoR)', which combines the ability to trust AI when it is right with the ability to judge for oneself when AI is wrong.

For effective collaboration with AI, going beyond simply improving accuracy, users need to be able to assess AI's reliability according to each context and adjust their level of trust accordingly. Reflecting this shift, LG AI Research has embedded the values of critical thinking and verification into the design philosophy of ChatEXAONE.