문화간 인공지능 연구센터
문화간 인공지능 연구센터
Center for Intercultural AI (CIAI)
The Center for Intercultural AI at Korea University's School (soon to become College, in 2025) of Media & Communication was inaugurated in January of 2024. Uniquely positioned to spearhead an AI ethics and policy research by leveraging diverse faculty composition across the spectrum of economics, business, humanities, data science, media & communication, and journalism, while also drawing on its current and prospective collaborators in the industry, this center aims to serve as the nexus of transdisciplinary collaboration and proactively explore the ethical dimensions of AI from an intercultural perspective.
The Center’s emphasis on “inter” and not “trans” or “cross” cultural sensitivity in AI ethics is a deliberate one that stems from the founding members’ interest in acknowledging the importance of cultural communities and values as the cornerstone of diversity, which we see as an imperative for the present and future of nonhuman intelligence. Whereas current trends in AI ethics research revolve around issues that directly impact the growing rifts between racialized, gendered, sexualized, and physically conditioned communities, endeavors to address the nuanced interactions between and across disparate national, linguistic, and cultural communities have been relatively sparse. This gap engenders the ironic situation wherein constituents of regions outside the Global North remain largely deprived of the agency to actively shape and impact the algorithmic agents that already form the base infrastructure of our media ecology, when in fact their habitat often serves as the sourcing ground for the materials, both in terms of raw data and low-level yet crucial components of human intervention in AI development and alignment. Given how recent breakthroughs in AI research and applications are predominantly driven by models that deal with cultural representations such as large language models and other forms of generative AI systems, we believe that inscribing cultural sensitivity as key indices and also technical parameters within the cognitive structure of AI models is one of the most urgent tasks in AI Ethics and Safety research.
Housed by a university that is home to leading scholars in the field of Korean (and East Asian) studies, and also drawing on the School of Media & Communication’s access to unique and culturally customized data from industry partners, Center for Intercultural AI aims to make tangible interventions by developing new alignment strategies through which cultural sensitivity, as an aspirational value in humane intelligence, could be incorporated into the ecology of our planetary habitat across both human and nonhuman agents.
The state of AI Ethics research within the disciplinary setting of the academia at present mainly centers on computer science and engineering rather than the Humanities and Social Sciences. Ethics, however, is an axiological domain that calls for socio-cultural and historical reflections in light of the relative and contextual nature of value systems. As one of the flagship academic institutions in a country that has struggled to identify its positionality within the throes of modernity, but also one that harbors the unique edge of wielding an exponentially increasing level of cultural appeal and technical prowess of late, the Center leverages geocultural liminality as a rich foundation on which to build a more inclusive terrain for technological innovation.
Conceptual Frameworks:
1. Intersectional Cultural Sensitivity as the Ethical Milestone of Alignment: Cultural sensitivity is a form of intelligence that is unique to the human kind in its dual acknowledgement of identificatory boundaries as both empirically existent yet phenomenologically fluid. Based on the belief that “ethics” relies on the advocacy of diversity as difference without giving way to value-weighted differentiation, we claim that there is a pressing need to address intercultural sensitivity as a key milestone in AI ethics.
2.Intercultural Approach as the Gateway to Humane Intelligence: Our goal is to build alignment strategies and datasets to address the socio-cultural nuances of cultures and languages that remain underrepresented in the current terrain of anglophone-centric AI ethics research. The pursuit of equitable diversity, we believe, is not an initiative that seeks to benefit the disenfranchised alone but also the foundation of humane intelligence in the most aspirational form.
3. Aspirational AI as Ambiguous Intelligence: We aim to encapsulate the human experience of both cognitive and affective ambivalence within AI models. Exploring AI models that produce disparate ethical/emotional and logical/rational responses, we seek to understand and replicate the complex decision-making process that humans undergo when faced with conflicting emotions and logic. This endeavor will more closely align AI's decision-making processes with human ambivalence. Human emotion is by nature non-binary, and this complex dynamic is often fueled and formed not only by individual experience or tendencies but cumulatively, repeatedly, pervasively, and persistently inculcated cultural customs that engender unconscious yet clearly existent communal intent. Based on the belief that incidents of less pronounced yet profoundly impactful breaches of emotional and by extension socio-cultural sensitivities must be addressed in acknowledgement of such ambivalence as an ethical issue, of which the persistent production of Orientalist and/or techno-Orientalist portrayals of East Asian cultures seen in generative AI model outputs serve as prime example, our goal is to develop alignment strategies that position affective and cognitive ambivalence as a key milestone of humane intelligence.
Director:
- Haerin Shin(Associate Professor)
Member:
- SunKyong Lee (Assistant Professor)
- Ahran Park (Assistant Professor)
- DamHee Kim (Assistant Professor)
- Sangwon Lee (Assistant Professor)
History
▶Establishment of 'Center for Intercultural AI' under the Information Culture Resarch Institute in March 2024
Location & Contact
서울특별시 성북구 안암로 145 고려대학교 미디어관 402호
02-3290-2257
- 간사 최훈민
ethernet@korea.ac.kr