28 August 2019
Panel on AI and Data Privacy Risk Management
This panel was presented by SGInnovate in partnership with N-CRiPT and NUS Smart Systems Institute. Moderated by Ms Suchitra Narayan (Director of Venture Building, SGInnovate), the panel comprised Professor Mohan Kankanhalli (Director, N-CRiPT), Professor Jussi Keppo (Principal Investigator, N-CRiPT) and Dr Bora Eang (Head of Data Science, AXA Singapore).
Most of the current AI techniques especially related to machine learning (ML) are heavily data-driven. For example, the recent success of neural networks based deep learning has been predicated upon availability of large-scale high quality data. While AI algorithms help unlock value from data, there are many attendant risks involved. They range from private sensitive data being leaked (directly or through inference) to adversarial attacks on ML. This is of significant concern due to regulatory compliance issues due to acts such as PDPA and GDPR. Other risks involve bias in the data as well as algorithms and fairness of outcomes. The speakers will first introduce the risks involved related to AI and Data Privacy. They will also touch upon the current thinking as well as the research being done to help manage these risk.