Prof Liming Zhu delivered his seminar 'Responsible AI Engineering: A System-Level Approach to Testing Transparency and Accountability' at AAII on 15 May 2024.
Topic: Responsible AI Engineering: A System-Level Approach to Testing Transparency and Accountability
Abstract: This presentation explores the systemic challenges and solutions in responsible and safe AI, emphasizing a system-level engineering approach to testing, transparency, and accountability. It covers the operationalization of high-level principles and frameworks through measurement science, system design patterns, and system-level understanding, without necessarily perfecting the underlying AI models. The talk will also touch on the latest developments in Australia's AI safety standards, the international AI safety summit, and regulatory movements.