Minister Audrey Tang of the Ministry of Digital Affairs (moda) attended the AI Summit London 2023 during London Tech Week on the 15th. She engaged in extensive discussions with Laurence Moroney, the Advocacy Lead for AI at Google, to explore AI democratization strategies that promote AI innovation and safety through broad public participation.

The theme of the panel discussion at the AI Summit London was "Impulse vs Control: A reality check of what  generative AI means for Innovation, Ethics and cybersecurity in the AI space – how do we regulate fast moving tech?” " The conversation was chaired by Natalia Modjeska, Research Director, AI & Intelligent Automationat Omdia – a UK technology ecosystem research institute.

During the conversation, Minister Tang shared that Taiwan had enacted legislation to address the harms caused by generative AI, such as deepfake fraud, synthetic pornography, and election manipulation, all of which undermine societal trust. The National Institute of Cyber Security, under the supervision of the moda, is expected to establish an "AI Evaluation Center'' by the end of the year. It will work closely with the National Science and Technology Council's "Trustworthy AI Dialogue Engine (TAIDE)" to promote open-licensed foundational language models that can adjust their interaction modes based on the actual application domains of different ethnic languages, thereby co-creating AI applications that meet the requirements of democratization.

In addition, Minister Tang participated in the AI Roundtable, with participants including Louise Sheridan, Deputy Director of the Centre for Data Ethics and Innovation (CDEI) in the Department for Science, Innovation and Technology (DSIT), UK, as well as members of the Office for Artificial Intelligence. The main topic of discussion was "AI Risks and building social trust policy."

Minister Tang emphasized that the potential risks of AI, similar to global issues such as pandemics and nuclear warfare, should be treated as a top priority and addressed collectively by global stakeholders. The development of AI should not only reflect the perspectives of developers or specific groups, but ought also to be guided by the "Taiwan Model" of openness, fairness, and transparency to ensure that everyone has equal access to, and use of, AI.

Minister Tang pointed out that the "Taiwan Model'' consists of two pillars: an open model and democratic alignment. The moda organized the Ideathon this year, which encourages public participation in the AI development process so as to collectively tackle challenges and achieve value alignment. In July of this year, Taiwan will collaborate with international teams to launch deliberative workshops through the "Future Democratization of AI" dialogue, starting from Taiwan and connecting globally to establish a consensus on the boundaries of AI. Finally, Minister Tang emphasized the necessity of formulating a secure and sustainable path for AI governance. This can be achieved by amplifying diverse voices and implementing the Taiwan Model to realize AI democratization and create an infinite future.