The government has established a principle that even when artificial intelligence (AI) is used in national research and development (R&D), final responsibility and rights over research outcomes belong to the researcher. It also set standards requiring researchers to disclose AI models used and their scope of application, and prohibiting the input of materials for national R&D project evaluations or paper reviews into external AI services.
The Ministry of Science and ICT announced on the 20th that it finalized the "National R&D AI Research Ethics Guidelines" following deliberation and approval by the operating committee of the National Council for Science and Technology Advisory.
This measure was taken as generative AI usage rapidly spreads in research settings, boosting productivity while raising new research ethics issues such as inaccurate information generated by AI, bias in existing data, security and personal information leaks, and citation of sources with unclear origins.
The basic principle of the guidelines is to limit the role of AI in national R&D to that of a "tool." Even if researchers use AI to collect or analyze and translate materials and write research results, responsibility and rights over the outcomes belong not to AI but to the researcher.
To this end, the government presented seven recommendations: verification of facts, independent judgment, ensuring transparency, managing result reliability, compliance with policies and regulations, securing security, and protecting personal information.
Researchers must personally verify whether AI-generated content contains errors, whether key materials are missing, and whether sources actually exist and are trustworthy. Instead of using AI-created materials as-is, researchers are required to review them based on logical validity and expertise, and reconstruct them by reflecting their own perspectives.
Standards for disclosing the use of AI have also been established. If AI was used during the research process, researchers must disclose not only the model and version but also the purpose and scope of its application, including information collection, data inference, organization of research data, translation, etc. The Ministry of Science and ICT provides a "Disclosure Form for AI Usage" and a self-checklist for researchers to confirm their compliance.
Security standards have also been specified. Researchers must use security environments and AI tools approved by their affiliated R&D institutions and take care not to input materials containing personal information into AI services. In particular, during national R&D project evaluations and paper review processes, evaluation target materials must not be input into external AI systems. Evaluators must not attempt to bypass legitimate review criteria or distort evaluation results using methods such as "hidden prompts" generated with AI.
Hong Soon-jeong, Director of the Performance Evaluation Policy Bureau at the Ministry of Science and ICT, stated, "AI is a useful tool that can innovatively improve research efficiency, but final responsibility always lies with the researcher." She added, "We hope this will serve as an opportunity to establish a culture of responsible AI use and ensure research integrity in national R&D."
in TECH
