2022년 한국 정부가 'AI 디지털 교과서' 도입을 발표했을 때, 이를 둘러싼 논의는 단순한 교육 정책을 넘어선 것이었다. "디지털 인재 100만 양성"이라는 구호, AI 교과서의 전격적 도입, 그리고 2024-2025년 사이 법적 제동과 전면 재검토—이 일련의 사건들은 교육 정책이 어떻게 특정 담론을 통해 형성되고, 내부 모순에 의해 균열되며,다시 재구성되는지를 생생하게 보여주는 사례다.
이화여대 이하나와 김종훈은 2026년 한국교육정치학회지에 발표한 연구에서 Fairclough의 비판적 담론 분석 방법론을 활용해 2022년부터 2025년까지 한국 AI 교육 정책 담론의 궤적을 추적했다.
분석 결과, 세 가지 핵심 발견이 제시됐다.
첫째, 정책 담론은 위기 서사를 통해 강력한 정당성을 확보했다. "디지털 대전환"이라는 위기감과 "디지털 인재 100만" 같은 경제주의적 목표, 그리고 막대한 예산 배정이라는 수치적 통치성 전략이 결합되어 정책에 거부하기 어려운 불가피성을 부여했다. 이 담론은 AI 교육을 단순한 교육적 선택이 아닌, 국가 생존의 문제로 프레이밍했다.
둘째, "개인 맞춤형 교육"과 "모두를 위한 교육"이라는 보편적 수사 이면에는 심각한 균열이 숨겨져 있었다. 학생 데이터의 파편화, 교사 위계화, 그리고 공교육 체계 밖 학습자들의 구조적 배제가 바로 그것이다. 겉으로는 포용적으로 보이는 정책 언어가 실제로는 특정 집단의 소외를 심화시킬 수 있음을 연구는 폭로한다.
셋째, AI 디지털 교과서의 법적 실패는 단순한 정책 수정이 아닌 전략적 재구성으로 이어졌다. 정책 책임은 개별 학교로 분산되었고, 경제주의적 담론은 "호모 심비우스"라는 인본주의적 공생 개념으로 재포장되었으며, 정책 대상은 전 생애주기로 확장되었다.
연구자들은 AI 교육이 중립적인 문제해결 도구가 아니라 권력이 작동하는 담론적 실천임을 강조한다. 기술 도입에 앞서 교육의 근본 가치에 대한 사회적 숙의와 윤리적 안전장치 마련이 선행되어야 한다는 것이다.
이 연구는 한국 사례를 중심으로 하지만, AI 교육 정책을 추진하는 모든 국가에 시사하는 바가 크다. 교사와 학부모는 AI 교육 정책이 등장할 때 그 뒤에 어떤 가치관과 이해관계가 작동하는지를 비판적으로 읽어낼 수 있어야 한다.
📖 *A Critical Analysis of the Formation and Reconstruction of Artificial IntelligenceEducation Policy Discourse* |
논문 원문
※ 이 기사는 교육학 연구를 바탕으로 작성되었습니다.
When South Korea announced a sweeping plan to introduce AI-powered digital textbooks into its public schools, it did so with a carefully constructed narrative: the country faced a digital transformation crisis, needed to produce one million digital talents, and only AI-integrated education could secure its future competitiveness. Then came the legal setbacks, the public controversy, and the quiet reframing of the entire initiative.
A 2026 study by Hana Lee and Jonghun Kim, published in the Korea Society of the Politics of Education journal, dissects this story through the lens of critical discourse analysis—and the findings reveal something far more complex than a policy success or failure. They reveal how power operates through educational language.
Tracing the Discourse
Using Fairclough's critical discourse analysis framework, the researchers traced the trajectory of South Korea's AI education policy discourse from 2022 to 2025, identifying how it was formed, where it fractured, and how it was reconstructed.
Three key findings emerge.
Finding 1: Crisis Narratives and Economistic Goals Drove Legitimacy
The initial AI education policy discourse derived its authority from a sense of national urgency. Phrases like "digital transformation" framed inaction as existential risk. The goal of producing "one million digital talents" translated abstract policy into a quantifiable target. Massive budget allocations—the practice the authors term "numerical governmentality"—made the policy appear both serious and inevitable.
This framing served a political function: by constructing AI education as a national survival imperative rather than a pedagogical choice, it preemptively delegitimized dissent. Critics could be dismissed as obstructing national progress.
Finding 2: Universal Rhetoric Concealed Structural Fractures
The policy's language promised "personalized education" and "education for all"—inclusive and student-centered on its surface. But the researchers identify deep contradictions beneath this rhetoric. The push for AI-mediated education carried significant risks of student datafication—the reduction of learners to data points for algorithmic processing. It also involved a de facto hierarchization of teachers, with AI positioned as the authoritative instructional tool. Perhaps most troubling, students outside the formal public school system—in alternative settings or marginalized communities—risked systematic exclusion from AI-enhanced learning environments.
Finding 3: Legal Failure Prompted Strategic Recalibration
When the AI digital textbook initiative encountered legal obstacles in 2024 and 2025, the policy response was not retreat but reframing. Responsibility was redistributed downward to individual schools, reducing accountability at the policy level. The economistic discourse of digital talent production was softened into a humanistic vocabulary—the concept of "Homo Symbious," or humans living in symbiosis with technology. The policy's scope was simultaneously expanded to encompass lifelong learning, broadening its reach while diluting scrutiny of any specific intervention.
The Bigger Picture
Lee and Kim conclude that AI education policy is not a neutral instrument for solving educational problems. It is a discursive practice through which particular values, interests, and power relations operate—often invisibly. The shift from crisis rhetoric to humanistic rebranding, they argue, should not be mistaken for genuine philosophical change; it is a strategic adaptation to political pressure.
The researchers call for substantive social deliberation on the fundamental values of education before further technological implementation, along with the establishment of ethical safeguards that are meaningfully enforced rather than rhetorically invoked.
For Educators and Citizens
While the study focuses on South Korea, its implications are broadly relevant to any country accelerating AI integration in education. Teachers, parents, and policymakers should examine AI education initiatives not only for their technological merits but for the value assumptions and power dynamics they embed. Policy language that sounds universally beneficial may, on closer inspection, serve narrower interests—and critical literacy in reading that language is an essential civic skill.
📖 *A Critical Analysis of the Formation and Reconstruction of Artificial Intelligence (AI) Education Policy Discourse (Critical Discourse Analysis)* |
Original Paper
*This article is based on educational research.*
