(Research Report) AI Ethics in Practice


Posted by on September 9, 2021  /  0 Comments

As the field of artificial intelligence (AI) has advanced and AI solutions are being increasingly deployed, debates around the ethical use of AI have arisen. AI ethics, however, are difficult to study systematically. Firstly, there is often disagreement about the meanings of ethical principles such as “fairness”, “accountability,” and “explainability.” Secondly, principles that would be ideal to achieve in theory may not always be practically implementable or enforceable.

At a policy level, governments have outlined their goals for AI development, including ethical AI, through policy documents, and different countries are at varying stages of translating these visions into practice. While existing legal and regulatory frameworks can be brought to bear on AI ethics there are also gaps, and a dearth of case laws mean that rulings relating to many important ethical questions are yet to be made. Meanwhile, AI developers are searching for technical solutions to ethical problems, such as how to make their models fairer, and decision-making processes more explainable.

In this research, we focus on the challenges of implementing AI ethics principles in two Asia Pacific Nations: Singapore and India. We ask:

What are the ethical, policy, regulatory, and technical challenges of implementing AI ethics frameworks in Singapore and India? Specifically:

  1. What ethical principles are being advanced and advocated for in policy and policy-relevant documents from the governments of Singapore and India?
  2. How do the existing legal and regulatory landscapes of Singapore and India address these ethical principles, and what are the gaps?
  3. Through case studies, what are the challenges of implementing AI ethics principles through technical applications in the real world?

We undertake this research by analysing policy documents, legal and regulatory analysis, and analysing two case studies. Our goal is to extract insights and recommendations that will be of use to policymakers in addressing the implementation of AI ethics.

  Download PDF   Email

Leave a Reply

Your email address will not be published. Required fields are marked *

*

*

*