Why do you care about AI Existential Safety?

There are increasing real-world examples showing how the misuse of machine learning models leads to the spread of misinformation, racism, and sexism. For example, sexually explicit AI-generated images of Taylor Swift are being shared on Twitter (I wrote about it here https://tuananhbui89.github.io/blog/2024/erasing-concepts/). With the rapid advancement of AI models, the risks are becoming more realistic. While large companies, who are mainly responsible for developing these models, prioritize business goals over ensuring model safety. Therefore, I believe that the academic community, with less influence from business interests, should take on the role of ensuring the safety and trustworthiness of these models for society as a whole.

Please give at least one example of your research interests related to AI existential safety:

My recent research paper is about erasing undesirable concepts from text-to-image diffusion models. I have been working in the field of Trustworthy Machine Learning since 2019, my publication list can be found here https://tuananhbui89.github.io/publications/.