Brief Description: As AI technologies are now being used in many varying high risk application domains in India and the U.S., it is imperative that they be worthy of trust. But what are those applications and how are they similar and different in the contexts of the two countries? Are the possible harms and benefits of AI the same? Can AI be a force for empowering society and reducing inequities or will it lead to a dystopian future in both countries? What are the best applications of AI for social good in the context of India and the U.S.?

Brief Description: A principled approach to Trustworthy AI needs to revisit, and think beyond, standard outcome-based notions (e.g., demographic parity, equal opportunity) that are suited to fair classification and aligned with certain civil rights and labor laws in the US. Ethical behavior and trustworthiness are inherently human qualities whose definitions vary based on the underlying context. What is justice, in theory and practice, also varies based on the law of the land. Following the western notion of business, applications of trustworthy AI are often seen as exchange of goods and services with utilities for various stakeholders, supported by certain beliefs about labor, ownership, property, rights. This track will focus on understanding the above in different non-Western contexts, important challenges therein, and fundamental trade-offs between multiple desired properties in trustworthy AI.

Topic Description:   Data-driven modern machine learning and computer vision algorithms are utilized in building (biometrics based) secure identity management systems. These systems have shown superlative performance in several nation-level identity projects. However, challenges such as fairness, robustness against attacks, explainability, trustability, and privacy preserving are still largely unaddressed. In order to build a trustworthy secure society, handling these challenges in the identity management solutions is exceptionally important. This can be achieved via carefully crafting frameworks, building unbiased and robust algorithms, and conscientious implementations of systems with explainability and privacy features. This track will highlight the challenges and possible future directions to mitigate them.

Topic Description: Trustworthy AI can be developed in two distinct ways: (1) individual AI systems that have properties such as fairness, robustness, and explainability, and (2) the institution of AI as a higher-level concept that is transparent and certified. The first way is more suited for awakening trust among experts, whereas the second way is more suitable for the general public. In this Roundtable, we will focus on the second institutional track for trustworthy AI and discuss the methods that may be pursued. The discussion will consider both United States and India, as well as local and central contexts where there are varying amounts of existing AI regulations.

Brief Description: Federated Learning (FL) is transforming the machine learning ecosystem from “centralized over-the-cloud” to “decentralized across-users”, in order to strengthen users’ privacy, enable much better personalized AI services, and overcome regulatory barriers in large-scale deployment. The goal of this panel is to discuss, shed light, and make recommendations on broad societal and technical fundamental challenges that would arise in FL, in particular on its trustworthiness, reliability, fairness and bias, security, applications, and regulatory aspects.