Who is Responsible for Responsible AI?
Pascale Fung, Chair Professor at the Department of Electronic & Computer Engineering at The Hong Kong University of Science & Technology (HKUST)
Following the publication of ethical AI principles and guidelines in various jurisdictions around the world, the challenge now is how to interpret them and how to implement these principles into our AI systems, applications, and usage. Academic researchers can play a significant role in helping translate principles, laws and regulations into actionable steps, and in proposing algorithms to implement them. More importantly, academic researchers can provide the thought leadership in terms of responsible AI design and thinking. For decades, the development of AI has been optimizing for engineering metrics such as accuracy, precision and recall. Responsible AI means a new paradigm of design, implementation and measure AI systems using additional metrics such as robustness, safety, fairness, privacy-preserving and others based on human values. On one hand, these metrics pose additional challenges to AI researchers; on the other hand, they provide new opportunities for innovation. In this talk, I argue that there is no other way of doing AI than responsible AI and present case studies on traditional NLP tasks that have been reinterpreted in this new paradigm to align with human values.
AI Ethics for All: A Broader Perspective on AI Education
Casey Fiesler, Associate Professor of Information Science at University of Colorado Boulder
Considering the constant stream of new controversies, it is no longer in question that concepts like ethics, responsibility, and justice are relevant to AI practice and research. However, these topics are still often framed as specialties rather than as foundational for every aspect of the field. This talk provides an overview of the current state of ethics education in computing and make a case for an integrated approach that embeds ethics and justice throughout as integral to the technical practice of computing. Moreover, I will argue for greater inclusivity in the field of AI and beyond, including how giving more people the knowledge and tools to be critical of AI will improve the state of the field for everyone.
Casey’s Presentation (8 mb PDF)
Trustworthy & Responsible AI Considerations: Principles | Policies | Practice
Alpesh Shah, Senior Director of Global Business Strategy & Intelligence at the IEEE Standards Association
The landscape of algorithmic systems and AI Ethics continues to evolve as algorithmic-oriented policies emerge, making scalable implementations of Responsible and Trustworthy AI systems challenging. In this talk, the presenter will share observations and insights on market responses and the soft norms and tools that the IEEE and its ecosystem have available to support the realization of the Rome Call for AI Ethics.
Alpesh’s Presentation (4 mb PDF)
Ethics-based auditing of AI - What it is and why it matters
Luciano Floridi, Professor of Philosophy and Ethics of Information, Oxford Internet Institute, University of Oxford