[Special Series: The Great Shift in Education Paradigms in the AI Era] Part 3. [The Gap] The Trap of Intellectual Outsourcing: AI Literacy and a New Educational Inequality

The spread of artificial intelligence is accelerating the "second- and third-level divides," which are determined by how and in what context technology is used, going beyond the "first-level digital divide," which was once defined by whether people owned devices. If access to computers or the internet was once the main source of inequality, the divide in the era of generative AI arises from the quality of use: whether people use AI as a partner in intellectual growth or simply consume it as a stand-in.

The spread of artificial intelligence is accelerating the "second- and third-level divides," which are determined by how and in what context technology is used, going beyond the "first-level digital divide," which was once defined by whether people owned devices. If access to computers or the internet was once the main source of inequality, the divide in the era of generative AI arises from the quality of use: whether people use AI as a partner in intellectual growth or simply consume it as a stand-in.

Applying the theory of "cultural capital" by French sociologist Pierre Bourdieu to today's classrooms, the intellectual environment and critical resources built up at home and at school make a decisive difference in how AI tools are used. According to an analysis of learning behavior released by a joint research team from Stanford University and MIT, groups of students from higher socioeconomic backgrounds and with high academic achievement used AI as a "sparring partner" to test and expand their own ideas. In contrast, students in environments with weaker educational support tended to engage in "passive consumption," copying answers provided by AI without verification and submitting them as assignments.

Ultimately, technology access alone does not close the education gap. It produces a paradox in which "intellectual polarization" deepens between those with the critical resources to actively control AI tools and those without them.

The Trap of "Intellectual Outsourcing" and the Decline of Higher-Order Thinking

The most serious side effect of overreliance on AI is "intellectual outsourcing," in which students' capacity for thought deteriorates. According to research on "cognitive offloading" by cognitive psychologists Evan F. Risko and Sam Gilbert, the more people rely on external tools for memory and reasoning when solving problems, the less effort the brain puts into activating the relevant cognitive pathways.

A moderate level of cognitive struggle and trial and error ("productive struggle") is essential for forming long-term memory and developing higher-order thinking. But when students hand over the entire thinking process to AI, from drafting an outline to structuring an argument to completing sentences, they lose the "thinking muscle" that lets them build logic and correct errors on their own. The result is a deceptive state of learning in which the process of internalizing knowledge is skipped and students are left holding only "finished-product knowledge."

This trap of intellectual outsourcing works differently across social classes and creates a new form of power structure. The warning that society could be reshuffled into a small group of "intellectual supervisors," who can design AI algorithms and critically evaluate their output, and a large group of "intellectual subordinates," who accept AI-generated information uncritically, is the most serious ethical challenge facing the education community.

Redefining AI Literacy: Critical Verification and Self-Directed Questioning

To overcome the risks of intellectual outsourcing and ease inequality, it is urgent to establish "critical AI literacy" with many dimensions, going beyond conventional information and communication technology education. UNESCO and the European Union's Digital Competence Framework (DigComp 2.2) redefine the AI literacy that future generations need not as mere prompt engineering, but as the ability to exercise critical, ethical and self-directed control.

AI literacy in the true sense consists of the following four key dimensions.

• Understanding the technical principles: The ability to clearly recognize the limits of AI, which does not reason deductively but is a large language model that predicts the next word based on statistical probability.
• Critical cross-checking: A sharp eye that questions AI output from multiple angles, including hallucinations, data bias and the legitimacy of cited sources, and traces the facts.
• Ethical judgment: The attitude of recognizing and weighing the ethical dilemmas hidden behind the technology, such as copyright infringement, privacy violations and environmental costs arising in the data collection process.
• Self-directed questioning (epistemic agency): The ability not to be led by AI's answers, but to adjust the level of one's questions to fit one's own purpose of inquiry and to retain agency over the final judgment.

AI literacy should now be treated as more than a matter of an information gap. It should be treated as an essential "basic democratic right," necessary for taking part in public discourse as an independent citizen and protecting intellectual autonomy.

The Role of Public Education and Policy Alternatives to Close the Education Gap

To prevent the intellectual polarization AI could bring and to build an inclusive education system, active institutional intervention in public education is essential. If the ability to use AI is left to be determined by the private tutoring market or individual circumstances, the ladder of social mobility through education will inevitably be shattered.

The public education system must first build a "public AI education infrastructure" that guarantees all students equal access to paid, high-performance AI tools and a secure data environment. It must also institutionally prevent the "AI cliff," in which access splits between lower-performing basic models and advanced premium models depending on economic circumstances.

In addition, the curriculum needs a sweeping overhaul that goes beyond simply adding an AI subject, integrating critical AI literacy education into every subject. Teachers should receive intensive training in "gap-closing AI pedagogy," which covers not only technical skills but also how to prevent students' intellectual outsourcing and foster independent thinking. Only when the direction of technology is guided by the values of public education, rather than leaving education to the advance of technology, can every student grow into a leading figure of the AI era.

© NY&NJ (뉴욕앤뉴저지). Unauthorized reproduction and redistribution prohibited

ⓒ New York and New Jersey. Unauthorized reproduction and redistribution are prohibited.