AI, Affective Robots and Ethics

By Laurence Devillers

Abstract

Is artificial intelligence (AI) developed according to ethical principles? After presenting herself as well as her research center, Laurence Devillers defines AI and explains how it works. She also presents the risks of the lack of women in the AI industry. She then gives a brief summary of AI historical development while focusing on the specificities of contemporary AI. To show the current limits of AI, Devillers enumerates the differences between machine intelligence and human intelligence. She insists on the fact that the ability for electronic systems to fake emotions can deceive vulnerable people. She then lists a series of ethical principles that should be respected by AI industry leaders. As a conclusion to her lecture, Devillers claims that there are no ethical machines. Ethics should come from the community of researchers and industries by building common AI tools.

Summary

00:00:00 – Introduction

00:00:10 – 1. Contextualization
00:00:10 – 1.1. Presentation of Laurence Devillers
00:09:42 – 1.2. Presentation of LIMSI-CNRS
00:10:06 – 1.3. Presentation of DATAIA Institute

00:14:42 – 2. Trustable AI
00:14:42 – 2.1. Definition of a trustable AI
00:16:12 – 2.2. Building a trustable AI

00:17:33 – 3. Women in AI
00:17:36 – 3.1. Statistics and issues

00:19:18 – 4. AI and robotics
00:20:48 – 4.1. A brief history of AI
00:23:16 – 4.2. Moral and technical consequences of automation
00:23:35 – 4.3. Contemporary AI

00:32:25 – 5. Machine vs. human intelligence
00:33:02 – 5.1. Machine vs. human perception
00:35:55 – 5.2. A narrow AI
00:38:00 – 5.3. New frontiers for AI

00:39:32 – 6. Human-machine co-evolution: issues and recommendations
00:40:26 – 6.1. Chatbot and AI
00:43:00 – 6.2. Emotions in social interaction

00:54:55 – 7. Ethical principles and industry leaders
00:55:00 – 7.1. Anthropomorphism
00:58:20 – 7.2. Bio-mimicry
01:00:10 – 7.3. Adaptive learning
01:01:33 – 7.4. AI “own” emotions
01:02:21 – 7.5. Intimate systems
01:04:15 – 7.6. Systems across cultures
01:04:40 – 7.7. Nudging

01:07:20 – 8. Partnerships to build an international HUB on AI
01:07:25 – 8.1. Ethical issues
01:10:00 – 8.2. AI for all

01:14:08 – Conclusion