| Generative AI |
AI that creates outputs such as text, sound, images, and videos by mimicking the structure and characteristics of input data. Generative artificial intelligence under Article 2, Subparagraph 5 of the AI Basic Act |
| Hallucination |
A phenomenon in which AI plausibly generates content that differs from facts or lacks a factual basis |
| Instruction (Prompt) |
The request and context given to AI. This book explains it through five elements: goal, audience, format, example, and constraint |
| Example-Based Instruction |
A method of guiding format and quality level by including examples of the desired output in the prompt. It increases accuracy and consistency with a few examples |
| Achievement Standards |
Statements describing the knowledge, skills, and attitudes that students are expected to achieve through learning in each subject. They can be found at the National Curriculum Information Center (NCIC) |
| Source Data |
Original materials serving as the source of facts. Public statistics, original texts of statutes, curriculum documents, historical records, etc. |
| Risk-Based Verification |
An inspection method that prioritizes verifying information whose potential harm is greatest if incorrect |
| Stereotype |
A rigid, preconceived notion about a specific group. It easily appears in character settings and role portrayals in AI-generated materials |
| Three Categories of Bias |
Systematic bias, computational and statistical bias, and human cognitive bias, as categorized by the NIST AI RMF |
| Differentiated Materials |
Materials that provide the same learning objective through different expressions and scaffolding tailored to student levels |
| Disclosed Work |
A work made available to the public. Subject to the provisions on use for instructional purposes and quotation under the Copyright Act |
| Work |
A creative production that expresses human thoughts or emotions (Article 2, Subparagraph 1 of the Copyright Act) |
| Use for Instructional Purposes |
Reproduction, public transmission, etc., of parts of disclosed works for instructional purposes by designated schools and institutions pursuant to Article 25, Paragraph 3 of the Copyright Act |
| Compensation |
Money paid to the holder of author's economic rights for use for instructional purposes. It is not paid for reproduction, etc., at schools of high school level or lower (Article 25, Paragraph 6) |
| Public Transmission |
Transmitting a work via wireless or wire communications so that the public may receive it or have access to it. Posting on a learning platform is a representative example |
| Quotation |
Using another person's work within a reasonable scope in accordance with fair practices for news reporting, criticism, education, research, etc. (Article 28) |
| Indication of Sources |
The obligation to indicate sources when using a work (Article 37). It also applies to use for instructional purposes |
| High-Impact Artificial Intelligence |
AI that has a significant impact on or is likely to cause risk to human life, physical safety, and fundamental rights, and is utilized in areas enumerated by the act (Article 2, Subparagraph 4 of the AI Basic Act) |
| Student Assessment Area |
Among high-impact areas, student assessment in early childhood, primary, and secondary education pursuant to Article 9, Paragraph 1 of the Framework Act on Education (Article 2, Subparagraph 4, Item (j)) |
| Artificial Intelligence Development Business Entity |
A person who develops and provides AI (Article 2, Subparagraph 7, Item (a)) |
| Artificial Intelligence Utilizing Business Entity |
A person who provides products or services using AI provided by a development business entity (Article 2, Subparagraph 7, Item (b)) |
| Affected Person |
A person whose life, physical safety, and fundamental rights are significantly affected by AI products or services (Article 2, Subparagraph 9) |
| Prior Review |
The obligation of a business entity to review in advance whether AI qualifies as high-impact (Article 33, Paragraph 1) |
| Request for Confirmation |
Requesting the Minister of Science and ICT to confirm whether AI qualifies as high-impact. It is an option, not an obligation |
| Human Management and Supervision |
One of the measures required of high-impact AI business entities. Enabling humans to review and correct AI results (Article 34, Paragraph 1, Subparagraph 4) |
| Explanation Measures |
Measures to explain final results, major criteria, and an outline of training data to the extent technically feasible (Article 34, Paragraph 1, Subparagraph 2) |
| Impact Assessment |
Assessing in advance the impact of high-impact AI on human fundamental rights. A duty of best efforts (Article 35) |
| Prior Notice |
Notifying in advance the fact that a product or service operates based on high-impact or generative AI (Article 31, Paragraph 1) |
| Indication on Outputs |
Indicating the fact of generation on generative AI outputs (Article 31, Paragraph 2) |
| Machine-Readable Indication |
An indication via a method readable by machines rather than humans (watermarks, etc.). When used, the fact of generation must be provided at least once through guidance text, voice, etc. (Enforcement Decree, Article 23, Paragraph 2) |
| Automation Bias |
The tendency of humans to excessively trust and follow the judgment of automated systems |
| Length Bias |
The tendency of a scoring tool to award higher scores to longer responses regardless of content |
| Inconsistency |
The property of AI producing different results each time for the same input |
| Prompt Injection |
An attack in which hidden instructions within the input alter the model's behavior. One conducted via external data is indirect prompt injection |
| Agreement Check |
Checking how identical AI scores are to teacher scores and where they differ by comparing them |
| Rubric (Scoring Guide) |
A table organizing evaluation criteria and performance level standards |
| Anchor Responses (Anchors) |
Representative responses for each level agreed upon by human raters. Used to align interpretations between humans and AI |
| Formative Assessment |
Assessment designed to check understanding and provide feedback during the learning process. It is generally not reflected in final grades |
| Personal Information |
Information relating to a living individual that identifies such individual. Includes information that can identify an individual when easily combined with other information (Article 2, Subparagraph 1 of the Personal Information Protection Act (PIPA)) |
| Pseudonymization |
Deleting or replacing part of personal information so that a specific individual can no longer be identified without additional information |
| Anonymized Information |
Information that can no longer identify an individual even if other information is used, reasonably taking into account time, cost, and technology. The Personal Information Protection Act (PIPA) does not apply (Article 58-2) |
| Data Subject |
An individual who can be identified by the information processed and who is the subject of that information |
| Personal Information Controller |
An institution, corporation, organization, or individual that processes personal information to operate personal information files for business purposes |
| Minimum Processing |
The principle of collecting and processing only the minimum personal information necessary for the purpose (Article 3) |
| Out-of-Purpose Use |
Using personal information beyond the scope of the basis for collection. A separate legal basis, such as separate consent, is required (Article 18) |
| Consent of Legal Representative |
Consent from a guardian, etc., that must be obtained when consent is required for processing the personal information of a child under 14 years of age (Article 22-2) |
| Sensitive Information |
Information that is likely to noticeably infringe upon privacy, such as ideology, beliefs, health, etc. Separate consent or a statutory basis is required (Article 23) |
| Entrustment and Trustee |
Entrusting personal information processing work to another party, and the party entrusted with such work. Documentation, public disclosure, and education/supervision are required (Article 26) |
| Cross-Border Transfer |
Providing, entrusting the processing of, or storing personal information overseas. It can be done when the statutory requirements are met (Article 28-8) |
| Leak, etc. |
The loss, theft, leakage, forgery, alteration, or damage of personal information. Upon becoming aware, notification must be given to the data subject, and a report must be filed if the criteria prescribed by the Enforcement Decree are met (Article 34) |
| School Life Records |
Records of students' personal information, school registers, attendance, curriculum learning developmental status, etc., prepared and managed by the head of a school (Article 25 of the Elementary and Secondary Education Act) |
| Restriction on Provision of Student-Related Materials |
Provisions prohibiting the provision of school life records and health examination records to a third party without the consent of the student (or the student and guardian if a minor). Statutory exceptions exist, such as supervision/audit, higher-level school admissions selection, and de-identified statistics/research (Article 30-6 of the Elementary and Secondary Education Act) |
| Use for Training |
An AI service using user inputs for model training or improvement |
| Retention Period |
The period during which an AI service retains inputs and conversation logs. It is a separate condition from use for training |
| Institutional Contract Account |
An account for which schools or institutions have established data terms through a contract. Terms may differ from personal accounts from the same provider |
| AI Classifier (Detector) |
A tool that estimates the probability that text was written by AI. It is not reliable enough to serve as the basis for assessing students |
| Authorship |
Who wrote which parts of a text and the responsibility for its content |
| Permitted Levels |
The scope of AI use determined for each assignment. This book explains it in four levels, from 0 (no use) to 3 (unrestricted use) |
| AI Usage Log |
A record documenting the tools, dates, purposes, key questions, and utilization methods used by a student |
| Explanation |
An individual under suspicion explaining their position and circumstances. A key step in academic integrity procedures |
| Data Manipulation |
Creating survey or experimental results that do not exist (fabrication) or arbitrarily altering actual results (falsification) |
| NIST AI RMF |
The AI Risk Management Framework from the US National Institute of Standards and Technology. Its core consists of four functions: Govern, Map, Measure, and Manage |
| Govern (Governance) |
The function of cultivating and implementing a risk management culture within an organization. A cross-cutting function across the other functions |
| Map / Measure / Manage |
The function of establishing the context of risk / The function of analyzing and tracking risk / The function of allocating resources to risk and responding to incidents |
| UNESCO Guidance for Generative AI |
The Guidance for Generative AI in Education and Research published by UNESCO in 2023 |
| UNESCO AI Competency Framework for Teachers |
An AI competency framework for a teacher consisting of five dimensions, 15 competencies, and three progression levels—Acquisition, Deepening, and Creation—published by UNESCO in 2024 |
| Pilot Operation |
Operating on a small scale first to identify effects and issues |
| Performance Indicator |
A metric used to determine whether the objective of adoption has been achieved |
| Knowledge Materials |
Materials used by a document-based tutor as the basis for its answers |
| Retrieval-Augmented Generation (RAG) |
A method of finding documents related to a question and generating answers based on their content. The standard architecture for knowledge-material-based tutors |