AI And Critical Thinking

Examining The Implications Of Letting AI “Think” For Us

Jul 01, 2025

AI offers incredible convenience. It can seemingly do everything while reducing the time and effort needed. However, we need to consider the true cost of defaulting to AI for more and more tasks. As artificial intelligence becomes an increasingly viable substitute for human intelligence, we risk weakening our capacity for critical thinking. This skill has long been the cornerstone of human progress. The time and effort we spend processing information help us uncover some of our most powerful insights. We should not be aiming for a future where AI does everything for us. We should be aiming for one where AI allows us to channel our energy into uniquely human tasks. Enhancing our capabilities, instead of diminishing them.

Why Critical Thinking Matters

Critical thinking is necessary for extracting the right insights.

Critical thinking is the ability to analyze, evaluate, and synthesize information. It pushes us to seek proof, question logic, challenge assumptions, identify contradictions, and spot falsehoods. The process of grappling with information is how we extract and form insights needed to make informed decisions. The lens through which we look at information influences what we find. Our context influences our thought process: what we think matters, what questions we ask, what information we seek out, etc. When we delegate a task to AI, we are letting it decide what to do, what to prioritize, what to focus on, etc. It’s going to look at the facts and arrive at a conclusion, which may or may not be the same as ours. Sometimes, it’s going to do the job just as well as we would (if not better). Sometimes, it’s going to do it worse. AI tools can certainly help us get an answer much faster than we would otherwise. However, it may not be the answer we really need.

Getting the ideas right means developing them well — drawing the conclusions that matter most, and exploring each one to the right level of detail. So getting the ideas right is not just a matter of saying true things, but saying the right true things.
— Paul Graham, Cofounder of Y Combinator

AI will not look for the same things we would because it operates using a different context. One that is a generalized “average” of all the information it’s trained on. When we use AI, we completely circumvent the work of deeply engaging with information. It simply won’t make the same connections we would when reviewing information or completing a task. Even if AI could give us seemingly “perfect” answers, we need to consider what it means to delegate the process of finding those answers.

I'm interested in the things I will see that other people would not have seen. And I think AI typically sees what everybody else would see.
— Ezra Klein, Journalist & Founder of Vox

Prompting is not the same as critical thinking.

Right now, there’s a lot of interest in prompt engineering. People are not yet familiar with how to converse with AI tools to get what they want. So, they are trying to figure out the “right” prompts to enhance and optimize the quality of AI outputs. A lot of “prompt engineering” is just having a well-structured conversation with AI. It’s about providing good context and constructive feedback. While this is certainly a skill that requires critical thinking, it is not a substitute for it. You don’t get the answers you need unless you ask the right questions, and you can’t ask the right questions unless you have the right information. When we let AI process information for us, we miss the opportunity to deeply engage with it. AI tools are not (yet) skilled at asking us the right questions to figure out (or help us figure out) what we want. So, even the best prompts might not get us the outputs we are looking for.

The Parallels Between Search Engines And AI Adoption.

The arrival of search engines transformed how we engage with information. When search engines first became popular, there were fears that the increased availability of “easy” answers would compromise our capacity for critical thinking. However, while search engines made finding information much easier, we still needed to analyze, evaluate, and synthesize information. It sped up the research process, allowing us to spend less time on information gathering, but it still required critical thinking, albeit a different kind from what we did pre-Internet. We had to develop the skill to process the raw information we found (from search results) into something meaningful, useful, and actionable. You could argue that we are once again in a similar position with AI. It presents the opportunity to drastically reshape how we do things because it can both think and act on our behalf. While the value of intelligent automation is undeniable, we need to consider where and how we implement it. There are real consequences for allowing AI to do our ”thinking” for us.

You May Also Like: The Future of Search

How AI Can Diminish Critical Thinking

Excess reliance on AI inadvertently contributes to cognitive atrophy.

AI’s major value proposition is its ability to increase efficiency and reduce cognitive load, which is the amount of mental effort we exert while thinking. AI can just help us get more done in less time, with less physical and mental effort. According to cognitive load theory, human beings have limited cognitive capacity. We can only retain and process a finite amount of information at any given moment. Therefore, reducing cognitive load enhances our performance because it reduces how much and how hard we have to think when completing tasks.

The perceived productivity and profitability gains have resulted in many companies actively trying to replace employees with AI. However, these benefits come at a cost. Recent research indicates that while AI can reduce immediate cognitive load, it may diminish problem-solving and critical thinking capabilities. The convenience of contextualized, personalized responses also reduces our motivation to independently research and understand information. For example, why would you read a lengthy document when you can just upload it to an AI tool and ask for a summary or ask it about specific topics?

However, letting the AI do more (if not most) of the thinking for us weakens our ability to concentrate and contemplate. People who complete tasks independently (without AI) must heavily engage their own cognitive resources, exerting greater mental effort and focus. In contrast, people who use AI tools must only lightly engage their cognitive resources. Essentially, they offload the mental effort and “thinking” to the AI. The less often we use our critical thinking skills, the less we stimulate the associated neural circuits, causing them to degrade over time. So, when it comes to critical thinking, we either "Use It or Lose It." The more dependent we become on using AI to complete a task, the worse we get at completing it without it.

We Still Need Critical Thinking When Using AI

We need to develop judgment to effectively assess and use AI outputs.

What people often fail to recognize and remember is that AI doesn’t interpret information the same way we do. AI outputs are probabilistic. They are the result of statistical guessing, not reasoning (at least not human reasoning). They make predictions about what a “good” response might be. Therefore, they can include inaccurate (or even totally false) information (hallucinations). What’s worse is that they can make their bad responses sound convincing by citing non-existent sources, presenting made-up explanations, etc.

It can also be hard to know (and judge) when AI outputs are generated based on incorrect logic, biased data, or incomplete information because they can be "black boxes" where inputs mysteriously turn into outputs. So, we still need critical thinking to discern how reliable and useful AI outputs are. Without it, we would just be using AI to get the wrong answer or do the wrong things faster. While the problem of the unreliability of AI outputs might be fixed with time, we still need to sharpen our critical thinking to leverage AI effectively.

Integrating AI into daily life requires balance. Its benefits are too great to ignore, but we can’t unthinkingly embrace them.
— Dr. Michael Gerlich, Center for Strategic Corporate Foresight and Sustainability

You May Also Like: Exploring Strategy For AI Products, AI Agents

Conclusion

Critical thinking will always be essential, but we might get worse at it if we are not careful.

AI is here to stay. Before we rush to automate everything with AI, we should consider what tasks should still be addressed primarily by people. These are things we (as opposed to AI) are best equipped to do. Where human thinking can yield better results than any algorithm ever could. Maybe we don’t fully know what that is yet because AI has such enormous potential. However, if AI is going to reduce the cognitive load involved in many (if not most) tasks, we need to figure out what to do with all the extra cognitive resources. We need to direct and invest our mental efforts in something to avoid wasting and diminishing our capacity for critical thinking.

Priank’s Newsletter

Discussion about this post