India Centre of Excellence in Information Ethics (ICEIE), UoH, has launched an International Consortium called Global AI Ethics Network for Social Good (GAIEN4SG) in 2021.

Under the Consortium, ICEIE has organised a talk on “How not to get entangled in ethical dilemmas when building AI systems – A practical vision for building trusted AI” by Ms. Padmashree Shagrithaya, Global Head-AI Analytics & Data Science, Capgemini, and Head AI @ Analytics – Business line, India.

Scholars from Russia, Ghana, South Africa, Israel, Italy, Romania, Spain, Jamaica and the Netherlands have participated.

“How NOT to get entangled in Ethical Dilemmas when building AI systems”-

Welcoming the participants, Prof. J. Prabhakar Rao, Director, ICEIE, and Secretary, GAIEN4SG has briefly outlined the objectives and the activities of the Consortium. He stated that the Consortium is an informal discussion platform to share research on AI Ethics and workout bi-lateral and multi-lateral academic collaborations in the area of AI Ethics.

The talk was chaired by Prof. Alfredo Ronchi, Politecnico di Milano, Italy, who specified the relevance of ethics in building and implementing responsible AI and welcomed the talk.

Ms. Padmashree in her talk focussed on seven key dilemmas of AI Ethics, such as automated decisions/AI Bias, Autonomous Machines, Ethical Data Sourcing, Exploitation with AI, Artificial General Intelligence (AGI), Societal Impact of AI Automation and Robot Ethics, and, finally, Sustainability.

She stressed that the ethical and moral issues of AI development and implementation have impact on financial and legal systems, planet and environment, human psychology, trust and society as a whole. She has also mentioned as to how to build a fair AI with privacy and data protection. She stressed the need for transparent and explainable AI with simulatability, decomposability and transparency, and local explainability. She was of the opinion that ethics are the basis to build a trusted AI by proposing 12 steps that can make AI Ethics into safe practice.

The talk has evoked wide discussion and the same is available on the link: