Speaker
Description
The so-called trolley problem is to ask AI for an ethical solution. However, such a problem is not unique to AI. A similar ethical question has also been there in China for many centuries : “Whom to save first when both your mother and your wife fall into water?”
Indeed, the ethics associated with AI are discussed at various levels. UNESCO and OECD are the leading international organizations that recognize the importance, implication and possible challenges of AI to human lives. The landscape of AI is very wide and diverse :
- From language translation to personalized content creation
- Scientific measurements and observations
- Audio and video generation, and music composition or painting
- Financial and legal assistance
- Healthcare and sophisticated medical detectors for diseases, cancer cells
- Chatbot, etc.
However, AI is not human. Thus, AI need not respect any ethics that are aimed for humans. But designers of AI can make AI obey human ethics. Thus, designers should be armed with “AI ethics” in order to make AI ethical, i.e., to make AI equipped with “ethics of AI.” These two concepts look identical, but have noticeably different aspects.
We all know that AI should be designed to ensure the users of the AI that :
-The whole life cycle of AI are transparent and explainable
-They [designers] are responsible and accountable for AI
-They have full awareness and literacy
-AI is designed, developed, set into operation with multi-stakeholder collaboration and adaptive governance
These guidelines may be called the AI ethics.
On the other hand, there is also the ethics of AI, which is to make AI ethical. AI should not help, suggest, decide or lead humans to immoral, unethical conclusions. It is definitely required when AI deals with subjects involving ethical values. Hence, ethics of AI should be installed or equipped beforehand in AI.
In this talk, I would discuss, including these two concepts, social responsibility in cyberspace.
