As AI-generated content becomes increasingly prevalent across creative industries, concerns about copyright infringement have come under the spotlight. Questions surrounding the use of copyrighted material in training Large Language Models (LLMs), and the difficulty in detecting unauthorized reproductions, have triggered widespread discussion among technologists, legal experts, and policymakers.
Addressing these timely issues, Professor Siu-Ming YIU, Associate Director (TPg) of the School of Computing and Data Science (CDS) at The University of Hong Kong, joined by Dr Jikun WU and Dr Li NING, both PhD graduates from the Department of Computer Science, were recently featured on TVB’s primetime programme News Magazine (《新聞透視》).
In the interview, the team discussed the technical complexities of identifying whether AI-generated works contain copyrighted material. They introduced their strategies for mitigating such risks, highlighting the importance of desensitization techniques to filter or remove sensitive copyrighted data during model training. Professor Yiu emphasized that while technological measures are critical, broader interdisciplinary collaboration—especially between the AI research community and legal experts—is essential to establish robust and transparent frameworks for the responsible development and use of AI-generated content.
The programme was broadcast on 19 April 2025 and is now available for viewing on the TVB website https://news.tvb.com/tc/programme/newsmagazine/68036203232aaaad609a2bce/%E6%96%B0%E8%81%9E%E9%80%8F%E8%A6%96-AI%E4%BE%B5%E6%AC%8A .