AI Governance Laboratory

In recent years , with the rapid development of artificial intelligence (AI) technology, AI has profoundly affected various aspects of our lives. Towards the end of 2022, the emergence of Generative Artificial Intelligence (Generative AI) along with large language models (LLMs) has brought about extensive applications in areas such as natural language processing, image generation, and AI manufacturing, ushering in a new era of AI development. However, the rapid advancement of AI also comes with significant challenges to human civilization, social norms, economic structures, and even the core values of democracy and the rule of law. Increasing societal polarization driven by algorithms, digital surveillance and targeted advertising, the spread of fake news, and the misuse of deepfake technology are some of the critical challenges. The primary goal of this laboratory's establishment is to address these challenges by returning AI to a human-centric approach and establishing trustworthy AI through AI governance, while responding to the potential impacts of AI technology on civilization and democratic society.

The establishment of the AI Governance Laboratory in our Institute of European and American Studies aims to conduct extensive and in-depth research on critical issues related to AI governance. This includes topics ranging from digital privacy, data protection and data governance, algorithm transparency and explainability, exploration of causality in automated decision-making, cognitive science and bias mitigation, human knowledge and wisdom reflection, human-robot interaction and responsibility allocation, as well as the establishment of AI ethics and norms. We hope to adopt a more open and diverse attitude, drawing on perspectives from various disciplines, including law, jurisprudence, cognitive philosophy, epistemology, ethics, and information science. Through interdisciplinary academic dialogues, we aim to establish horizontal connections to contemplate essential issues in AI governance.

Taking the governance of Generative AI as an example, the current EU AI Act still relies on a risk-based regulatory logic, which may not effectively address the unequal power dynamics of risk and responsibility in the application of large language models. Furthermore, the current EU governance mechanism has yet to consider the uniqueness of various global languages and their cultural contexts, as well as the relationships between context, narrative, and cultural metaphors. As a major user of traditional Chinese characters in the world, Taiwan holds significant importance in shaping international standards for Generative AI. The AI Governance Laboratory will actively engage in the governance of Generative AI, enriching global discussions on AI governance through a dialogue between global norms and local forms.

We hope that through this dual perspective, we can focus on two aspects. On one hand, we aim to engage in a dialogue with European and American ideas and AI governance norms, with a focus on the development of global AI regulations. On the other hand, we want to incorporate the specific characteristics of the local context to avoid the frameworks or limitations that may arise from a Euro-American-centric perspective. This research approach, by introducing different social contexts, actors, structures, and cultures, will create novel and original European and American studies through comparison, dialogue, and translation. These perspectives and dimensions will enrich traditional Euro-American studies, which are mainly focused on regional or area studies, making them more inclusive, original, and critical, and also fostering a more open and diverse discussion on AI governance.

Breaking free from conventional frameworks and imaginative experimentation is a crucial catalyst for achieving this dual perspective. The composition of this AI laboratory with members from diverse backgrounds is the fundamental condition for unleashing imagination. We hope that this laboratory can offer different perspectives on AI governance due to the diversity of its members. Ultimately, we hope to stimulate the creation of new ideas through the interaction and exploration of different interdisciplinary theories and provide academic analyses and recommendations for various AI governance issues in response to the urgent impact of AI technology on future society and human civilization.