LLMs and Critical Thinking

...or am I losing my mind?

I’ve been thinking about how I work with MAL in relation to recent conversations about critical thinking and chatbots.

When I use AI, I begin from my own thoughts about the work I’m doing – how it feels, what I think is happening, where I’m unsure. I don’t ask the machine to generate ideas for me. I provoke it. I push it to offer alternative readings, counter-positions, and different framings. I edit its responses, argue with them, and feed my disagreements back into the exchange. The point is not agreement, but pressure.

This makes thinking easier in one sense, but also more demanding. It opens directions I wouldn’t have reached alone, and then asks me to decide which of those directions matter. I remain responsible for the choices. The machine gives me something to think against.

Watching Advait Sarkar’s talk on AI and critical thinking sharpened this for me. His concern – that conversational AI can erode our capacity to think critically by outsourcing judgment – feels right. Where I hesitate is in the suggestion that removing the chat interface solves this problem. A missing chat box does not mean a missing conversational paradigm. If the structure remains “I write, the system responds, I read”, then the epistemic risk is still present, only less visible.

In some ways, hiding the dialogue may even increase the risk. When the back-and-forth is explicit, I can see influence happening. I can track revisions, tensions, disagreements. When AI becomes ambient, embedded, or invisible, it becomes easier to mistake machine-shaped thought for one’s own.

For me, the question is not how to eliminate conversation with AI, but how to stay awake inside it. The danger is not disagreement, but passivity. Not augmentation, but atrophy. Critical thinking survives when friction is maintained and when the human insists on being the one who chooses, edits, refuses, and lingers with uncertainty.

I don’t think this tension can be designed away. It has to be practiced.

I’m responding here to a talk by Advait Sarkar, How to Stop AI from Killing Your Critical Thinking (TED, 2024).

Advait Sarkar: How to stop AI from killing your critical thinking | TED Talk)