Welcome to our Fall PhD Colloquia Series, where we present the innovative research of our PhD researchers in the field of urban development. Each session explores a wide variety of topics, providing deeper insights into their academic research.
Somesh Sharma
Somesh Sharma is a Specialist in Environment and Sustainable Development Planning at IHS. His research, entitled "Explainable AI for Automated Vehicles: Addressing Liability in Accidents," explores the explainability of AI models used in automated driverless vehicles (ADVs). His work analyses how the "black box" nature of AI systems impacts accountability and public trust, particularly in the context of accidents involving ADVs.
More about his research
Somesh’s research aims to contribute to the development of regulatory frameworks that enhance the transparency and safety of ADVs in cities. Using qualitative data analysis of legal cases, he investigates how explainable AI (XAI) can inform policy decisions and improve the design of AI systems.
The main highlights of his research are:
- Human-machine interaction in ADVs: The research investigates how the interaction between humans and automated vehicles is changing traditional roles, focusing on the shift from active driving to passive passenger roles. This theme explores the social and behavioural implications of this change, especially in urban environments.
- Safety concerns in urban settings: Somesh analyses the unique safety challenges posed by ADVs in cities, including their reliance on external systems like GPS and Wi-Fi. The focus is ensuring that ADVs can operate safely in urban environments, accounting for technological limitations and human unpredictability.
- Legal challenges in ADV liability attribution: To examine how legal frameworks address the attribution of liability in cases of ADV accidents, Somesh looks at the difficulties in assigning responsibility when an AI system makes a decision and how new regulations are created to ensure clarity in accountability.
