Insights on trusted and safe AI

Speaker:
François Terrier – Director of Research at CEA – VP, Artificial Intelligence program
Director of research at CEA, François Terrier is responsible for the Artificial Intelligence program at CEA Tech and for the Trusted AI strategy carried by the CEA List. After a PhD in AI and 10 years of R&D in the field, he conducted research on the engineering of software and trusted systems. Head of department from 2010 to 2020, he was in charge of the development of activities on model-based engineering, formal methods, and of the emergence of new activities on cyber-security, resilient distributed systems, including blockchain, trusted and embedded artificial intelligence.

Abstract:
As artificial intelligence becomes more sophisticated, uses for the technology will proliferate and expand into totally new areas. AI has already made inroads into consumer goods, and industrial companies are now looking to the technology to bring major improvements to their products and systems. These opportunities are not without risk, however. AI raises serious questions about trust, sovereignty, and responsibility to people, property, and society. The European Union issued a white paper, On Artificial Intelligence – A European approach to excellence and trust, outlining the bloc’s approach to AI. For safe, secure AI to become a reality, data ownership and confidentiality must be protected. The French government is addressing these challenges with initiatives like AI for Humanity and the dedicated “Great challenge on trustworthly AI” supporting the launch of a very ambitious program, Confiance.ai, with national level major industrial players in the fields of Automotive, Aeronautics, Defense, Manufacturing and Energy.