평가결과

Overview



■ (Subject of Assessment) AI Video Synthesis Services

■ (Definition) Services that utilize deep learning technologies to analyze video data and generate new videos or modify/transform existing video, image, or audio content based on that analysis

■ (Rationale for Selection) Deepfake technology poses significant social risks, including the undermining of democratic values, defamation, and sexual exploitation. In recent years, the range of potential victims has expanded to include the general public, particularly children, adolescents, and women, highlighting the urgent need for effective responses.

- At the same time, exploring ways to strengthen its positive functions and ensure responsible use holds substantial potential to drive innovation and generate social benefits across diverse sectors such as content creation, education, and entertainment. Accordingly, a balanced assessment and prudent policy responses are required for this technology.

​

Key Findings of AI Ethics Impact Assessment

■ Impact Identification and Assessment

- Based on responses from an expert evaluation panel and a public forum group, the ethical impacts of AI video synthesis services were identified: 55 positive and 66 negative in total. Each identified impact was assessed for its magnitude, scope, probability of occurrence, and feasibility of resolution.*

* Methodological references: UNESCO’s (2023) "Ethical Impact Assessment: A Tool of the Recommendation on the Ethics of Artificial Intelligence," and the Ministry of Science and ICT and the Korea Information Society Development Institute (KISDI)’s (2024) "AI Ethics Impact Assessment Framework."

■ Analysis of the Needs for Government Policy Support and Responses

- (Areas Requiring Government Support)

① (Continuous Management: Areas where active and systematic government support is required to ensure the continuous creation of positive effects)

☞ Protecting socially vulnerable groups, incentivizing the development of privacy protection technologies, etc.

② (Public Welfare: Areas where positive effects may be limited but government-led action is required for public safety and welfare)

☞ Strengthening safety awareness, reducing social costs to promote solidarity, etc.

③ (Private Autonomy: Areas where the private sector can lead responses, with the government focusing on indirect support and guidance)

☞ Expanding accessibility to creative tools, invigorating creative activities, etc.

- (Areas Requiring Government Policy Responses)

① (Attention Required: Areas where the government is currently responding appropriately to concerns, but continued monitoring is still necessary)

☞ Criminal exploitation arising from security vulnerabilities and malicious use, proliferation of digital sexual crimes, etc.

② (Urgent Resolution: Areas where public concern over negative impacts is high but government response remains relatively limited, requiring priority intervention)

☞ Spread and dissemination of false information/fake news, widening intergroup disparities, psychological and physical harm to victims, etc.

■ Roles and Efforts of Key Stakeholders for the Ethical Development and Use of AI Video Synthesis Services

- (Businesses and Individuals) To establish a culture of ethical development and utilization of AI video synthesis services across society, the cooperation and voluntary efforts of all members of society-including enterprises and individuals-are essential.

· (Businesses) △ Developing technologies to enhance AI reliability and safety, △ Establishing responsible AI management and safety systems, △ Formulating corporate ethical responsibilities and trust-based policies, △ Expanding employee and consumer education and engagement, △ Strengthening external cooperation and communication

· (Individuals) △ Enhancing AI literacy and ethical awareness, △ Strengthening responsibility through active participation and practice, △ Protecting personal information and exercise rights responsibly, △ Promoting responsible creation and sustainable content consumption

- (Government) Systematic government support and responses are required to promote the sound development and responsible utilization of AI video synthesis services.

· (Policy Dimension) △ Strengthening AI education and literacy, △ Improving government regulation and management systems, △ Promoting industrial advancement and research support, △ Supporting global cooperation and standardization efforts

· (Legal and Institutional Dimension) △ Revising and improving laws and regulations, △ Preventing crimes and misuse/abuse, △ Regulating platforms and technology providers, △ Reinforcing the foundation for self-regulation through government support and policy guidance

· (Technological Dimension) △ Developing detection and verification technologies, △ Enhancing technical safety and ethics, △ Promoting R&D and technology investment, △ Fostering collaboration and exchange through government support and policy guidance