Date: October 10, 2025 (Friday)
Time: 15:30 – 17:00
Venue: Room 10.13, 10/F, Run Run Shaw Tower, Centennial Campus, HKU
Speaker:
Professor Maximilian Kiener
Institute for Ethics in Technology, Hamburg University of Technology
Abstract:
The scholarly literature presents various approaches to integrating ethical considerations into the development of AI, most prominently the approaches ‘ethics by design’ and ‘value sensitive design’. However, despite some clear strengths of these approaches, there are also two significant weaknesses. First, there is still insufficient technical engagement. Ethical discussions frequently gloss over the technical details necessary to be relevant for actual AI development. Second, there is a superficial treatment of ethics, such that stakeholder preferences are sometimes conflated with genuine normative principles. These two criticisms, though powerful on their own, reveal an even greater challenge when combined: the need for a more comprehensive understanding of how ethical values and principles intersect with technical requirements. This talk makes a start in addressing this combined challenge, and it analyses how to integrate ethics more effectively into the training of AI systems. In particular, I shall explore a new method of doing so, called Ethical Triangulation, which translates qualitative ethical considerations into the quantitative metrics needed for AI training by leveraging the interplay of three factors: abstract principles, context-specific judgments, and practical constraints. I shall illustrate this approach with examples from reinforcement learning and Markov Decision Processes. The key outcome is support for a new field of inquiry called ‘deep ethics’: ‘deep’ because it integrates ethical reasoning into the architecture of AI itself, and because it explores a foundational link between ethics and intelligence.
Don’t miss Prof. Kiener’s another talk at the HK Ethics Lab – Vagueness & Responsibility on October 8, 2025.
