Thinking for Ourselves in an Age of Ready-Made Answers
Obtaining an answer has become easy. The difficulty lies in knowing whether that answer deserves our trust, and whether we understand it ourselves. This distinction is central to Helen Pearson’s article on critical thinking in the 24 September 2026 issue of Nature. She describes how artificial intelligence (AI) can take over tasks that previously required students to investigate, compare and reason for themselves (Pearson, 2026).
Critical thinking is a form of judgement: the ability and willingness to assess what we have reason to believe and how we ought to act. AI makes this capacity particularly important because readily available answers can tempt us to bypass the inquiry that responsible judgement requires. We must still assess the reasons supporting those answers and the consequences of using them. From the perspective of practical philosophy, this concerns both knowledge and our relationships with other people.
Discerning and Judging
The word “critical” can be traced to the Greek κριτικός (kritikós), meaning capable of judging or discerning. It is related to the verb κρίνειν (krínein): to separate, distinguish, judge and decide (American Heritage Dictionary, n.d.). The history of the word offers a point of entry into the subject: critical thinking involves distinguishing between claims and their justifications, knowledge and assumptions, strong and weak arguments.
In everyday language, “critical” is often associated with finding fault. Yet someone who rejects everything does not necessarily think more carefully than someone who accepts everything. Both may avoid the work of assessing each particular claim. I therefore understand a critical stance as open to both justified doubt and justified trust. Inquiry must allow us to change our minds, including when this means acknowledging that someone else is right.
Practical philosophy sharpens this inquiry by distinguishing between empirical and normative questions. Whether an intervention works must be investigated through experience and relevant evidence. Whether it is just also requires consideration of values, interests and consequences. Even an effective intervention may be unfair or violate someone’s dignity. The transition from “this works” to “we ought to do this” therefore requires ethical justification.
In Aristotle, we encounter the concept of phronesis, commonly translated as practical wisdom. It concerns deliberation about how we ought to act in particular situations, with a view to a good human life (Aristotle, 2009, Book VI). Critical thinking and practical wisdom are distinct concepts, but they can support one another. By examining our reasons, we become better equipped to recognise what is at stake for those affected.
Consider a professional working with someone who repeatedly misses appointments. An immediate explanation might be that the person “lacks motivation”. This formulation appears to be a factual assumption about the cause of the absences. At the same time, it may convey a moral judgement: the person is failing to take responsibility or to do what they ought to do. An assumption about what is happening can thus readily become a judgement about who the person is.
Our judgement is tested when this explanation is examined. Have the appointment arrangements been understood? Is transport difficult? Do the meetings feel unsafe? Has the service made participation unnecessarily demanding? These questions concern factual circumstances, but they may also change how responsibility is understood. Perhaps the service needs to change its practices. Perhaps expectations need to be clarified with the person concerned. Even where motivation is limited, the ethical question of how the person ought to be treated remains.
Knowledge as a Basis for Judgement
Pearson emphasises that critical thinking requires subject knowledge. We may be skilled at assessing arguments in one field and considerably less capable in another. Someone investigating a historical claim needs to understand sources and context. Someone evaluating research needs insight into the field and its methods. A general exhortation to “think for yourself” does not provide this knowledge (Pearson, 2026).
Intellectual independence therefore also involves drawing on the insights of others. No one can verify everything alone. We need experts and institutions whose work can be scrutinised. A considered assessment also involves asking whom we have reason to trust and recognising the limits of our own expertise. “I know too little about this” can be a well-founded conclusion.
Critical thinking can be learned through engagement with concrete questions. A research review by Abrami and colleagues found positive effects from approaches including dialogue, authentic problems and mentoring (Abrami et al., 2015). These findings support combining subject knowledge with opportunities to investigate, discuss and receive feedback on one’s assessments.
From the perspective of practical philosophy, such conversations should allow participants to ask what makes an argument stronger or weaker. We can practise by examining a newspaper article together, discussing a difficult decision or explaining why we have changed our minds. In education, it is similarly reasonable to value the justifications and revisions through which an answer is reached. Learning then becomes more than the production of a conclusion.
Pearson also highlights metacognition: monitoring one’s own understanding and learning. Explaining a point in our own words, without the text in front of us, can reveal the difference between recognition and understanding. The question “What could make me change my mind?” directs inquiry towards our own position. Our preferred explanations must also be open to examination.
Trust in the Face of Disinformation
Misinformation may be passed on in good faith. Disinformation involves deliberate deception. However, we do not always know the intentions of the person communicating it. Before attributing a motive, we need to examine the content, sources and context.
When a claim makes us angry or highly enthusiastic, it may be wise to pause before sharing it. Perhaps it confirms what we already believe, provides a scapegoat or makes a complicated situation reassuringly simple. Emotions indicate that something matters to us, but they do not establish whether an explanation is sound. Critical thinking also involves examining information we would like to be true.
One useful approach is lateral reading. When we encounter an unfamiliar website, we leave it to investigate what other credible sources say about its publisher and its claims. We look for the original evidence and check whether several websites are relying on the same source. McGrew describes educational interventions that help people use such approaches (McGrew, 2024). An appealing presentation provides a weak basis for trust. Locating the original source allows us to examine whether the evidence actually supports the claim.
Pearson also discusses an approach in which people are shown examples of misleading information and the techniques used to produce it. She cites research suggesting that this can improve adolescents’ ability to identify unreliable social media posts (Pearson, 2026, p. 875). From the perspective of practical philosophy, this awareness should be combined with knowledge of the subject: recognising a rhetorical technique does not, by itself, determine whether a claim is true.
The critical assessment of sources also needs a purpose: to make justified trust possible. If all information is declared suspect, it becomes difficult to distinguish well-supported claims from unsubstantiated suggestions. A reasonable practice is to make our degree of trust proportionate to the quality of the justification. Some things we know with considerable confidence, some we know provisionally, and others we do not yet have sufficient grounds to assess.
AI and Our Habits
Can AI make us less inclined to think critically? This is a real possibility, and there is evidence that certain ways of using it can undermine learning. Rapid, well-formulated answers can make it tempting to bypass inquiry. We may also confuse access to an explanation with our own understanding of it.
An experiment in mathematics education illustrates this risk. Students given access to an AI tool with few instructional safeguards performed better while receiving assistance, but worse than the control group when that assistance was withdrawn. A version designed to support students’ own work reduced this risk (Bastani et al., 2025). The study provides evidence about learning under particular conditions. It does not establish how AI affects people’s general willingness to think critically.
Pearson describes a mixed body of research. AI tools are relatively new and varied, studies produce conflicting results, and too few rigorous or long-term investigations have been conducted. She also reports that some studies suggest AI can help people learn when the tools provide feedback or tasks that encourage independent thinking and problem-solving (Pearson, 2026, p. 874). This gives us reason to distinguish between different forms of use. The article does not warrant a general conclusion that AI diminishes people’s capacity to think.
The question for practical philosophy therefore concerns the habits we develop. If we repeatedly allow AI to formulate both the problem and the conclusion, we may have fewer opportunities to exercise our own judgement. If we use it to examine a provisional assessment, the conversation may give us further questions to explore.
A simple practice is to begin with a brief assessment of our own. Write down what you believe, why you believe it and where you are uncertain. This gives you something against which to evaluate the AI’s response. Then ask for a strong objection, an alternative explanation or help in identifying the assumptions underlying the argument. A question such as “What am I overlooking?” can open up the inquiry.
The next step is to verify what can be verified. Do the sources exist? Do they actually say what the AI claims? Does the conclusion follow from the evidence? AI can suggest useful questions for checking an argument, but its answers must themselves be examined. Repetition of a claim within the same conversation provides no independent confirmation.
Finally, offer an explanation in your own words, without copying the AI’s response. What do you now believe, and what, if anything, caused you to change your mind? Distinguish between what is now better supported and what remains unresolved. In this way, using AI can become an opportunity to exercise responsible judgement.
Making Room for Reflection
Critical thinking requires an environment in which people feel able to exercise it. Calling for independence achieves little if questions are met with irritation or changing one’s mind is regarded as a defeat. A teacher, leader or conversation partner can demonstrate that a reasoned objection is valuable, including when it challenges their own assessment.
Judgement should not be confused with quick and confident speech. Some people think best through writing; others need time before responding. A good learning environment offers several ways to participate. Encouraging critical thinking involves making room for pauses, provisional formulations and questions that have yet to take shape.
AI can remove much of the waiting time between a question and an answer. Yet if we also allow it to remove the work of formulating the problem, weighing reasons and engaging with objections, we risk losing precisely the opportunities through which judgement develops. Making room for reflection is therefore a deliberate choice about which parts of the work we will undertake ourselves.
We began with the distinction between obtaining an answer and knowing whether it deserves our trust. Critical thinking enables us to examine this distinction more carefully. When an answer is used to assess another person, share information or choose a course of action, we must also ask what values and consequences it entails. AI can assist us in this inquiry. We exercise judgement by working through the reasons ourselves.
References
Abrami, P. C., Bernard, R. M., Borokhovski, E., Waddington, D. I., Wade, C. A., & Persson, T. (2015). Strategies for teaching students to think critically: A meta-analysis. Review of Educational Research, 85(2), 275–314. https://doi.org/10.3102/0034654314551063
American Heritage Dictionary. (n.d.). Critic. In The American Heritage dictionary of the English language. https://ahdictionary.com/word/search.html?q=critic
Aristotle. (2009). The Nicomachean ethics (D. Ross, Trans.; L. Brown, Ed.). Oxford University Press.
Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), Article e2422633122. https://doi.org/10.1073/pnas.2422633122
McGrew, S. (2024). Teaching lateral reading: Interventions to help people read like fact checkers. Current Opinion in Psychology, 55, Article 101737. https://doi.org/10.1016/j.copsyc.2023.101737
Pearson, H. (2026). How to stay smart in the age of AI: The science of critical thinking. Nature, 657(8133), 873–875. https://doi.org/10.1038/d41586-026-02930-6
This essay was written in a conversation with ChatGPT









