Responsibility in the Times of Generative AI, seminar by Cristian Canton Ferrer
On 28 January 2025 Cristian Canton Ferrer, Head of Responsible AI at Meta, gave an invited talk on Responsibility in the times of Generative AI in the Auditorium of the Modena Technopole.
Abstract
As Generative AI technologies become more prevalent and used across more and more platforms, systems and applications, it is important to revisit how we apply principles of responsibility to them. Fairness, transparency, safety, robustness and other dimensions must be considered when developing GenAI algorithms, as well as keeping an ethics perspective. This seminar revisits the standard concepts of responsible AI and how they apply to GenAI, with a walkthrough of the Llama 3 model and the efforts to make it more responsible.
Speaker
Cristian Canton Ferrer leads the Responsible AI organization at Meta, covering fairness, robustness and safety, governance and accountability, and transparency, especially for Generative AI. Previously at Meta he led the computer vision team working on harmful content, misinformation and manipulated media. He was at Microsoft Research in Redmond and Cambridge (2012–2016) and at Vicon in Oxford (2009–2012), and received his PhD and MS from the Technical University of Catalonia.