At Cypher 2023, a distinguished panel of experts convened to delve into the complex subject of trust in Artificial Intelligence (AI). The panel was moderated by Raghavendra Rengaswamy, EY GDS Consulting Data and Analytics Leader, and featured Padmashree Shagrithaya, Executive Vice President & MD – Insights & Data GBL at Capgemini; Shubhosree Dasgupta, Head of DAO India at HSBC; Samarth Gupta, V.P Data & Analytics Engineering at Natwest Group; and Vikas Behrani, Vice President – Data Science | Artificial Intelligence and Machine Learning Practice at Genpact. The conversation covered a wide array of topics, from governance and accountability to ethics and societal values.
Governance and Accountability
The panelists, led by Raghavendra Rengaswamy, emphasized the importance of governance and accountability in AI systems. They discussed the need for robust frameworks to ensure responsible AI implementation. The concept of “Red Teams,” or ethical hackers, was also introduced as a means to test the security and reliability of AI systems.
Ethical considerations took center stage during the discussion. Padmashree Shagrithaya spoke about the need for an ethics board within organizations to oversee the responsible use of AI. The panelists also weighed in on the balance between risk and reward, emphasizing that while AI has enormous potential, it also comes with ethical and societal implications that must be carefully managed.
Data and Model Accuracy
The challenges of ensuring model accuracy were thoroughly examined. Padmashree Shagrithaya pointed out that the quest for the “highest level of accuracy” could be a myth and that organizations should focus on areas where accuracy is most crucial. She also introduced the concept of “liquid AI,” which aims to make AI models more adaptable and less data-intensive.
Human Values and Societal Impact
Samarth Gupta led the conversation on the alignment of AI systems with human and societal values. He discussed the challenges of defining what exactly constitutes human values in the context of AI and how these could be integrated into AI systems. He suggested that AI could either be constrained by human-defined values or be subject to a layer of human review to ensure ethical behavior.
The panel discussion at Cypher 2023, moderated by Raghavendra Rengaswamy, provided a comprehensive look into the complexities of trust in AI. From governance and ethics to data accuracy and societal impact, the conversation highlighted the multi-faceted challenges and considerations that come with the responsible use of AI.