Humans vs Robots: should we produce systems smarter than we are?
Artificial Intelligence is without doubt a hot topic these days. However, along with wide-spread interest and curiosity, popularity also brings confusion and misunderstanding. The purpose of the ‘Ethics in AI’ panel is to engage and shed light on some of the myths around AI. More specifically, the panel will focus on the issues of training AI systems and the consequences for humankind of creating Artifical Intelligences.
We will start by discussing how humans train AIs. What does the process entail and how difficult is it. We will dive into the topic of intelligent systems. If machines are the product of human making, is it ever possible that machines could out-smart us? Under what conditions is it possible or likely? This opens up questions of training biases and stereotypes and the possibility of their transference during AI training. Another important consideration is whether AIs can self-train. What could AI systems learn on their own and how? Could (or should) humans impose limitations on the robotic learning systems? What kind of implications could this bring to the field of AI? And then, is it possible that as a result of training, humans will end up being managed by machines? How likely is it that we will live in the age of Terminator?
Edward Pogossian (NAS RA)
Michel Daydé (IRIT)
Aghasi Poghosyan (PicsArt Inc)
Alexei Belotserkovsky (BASNET)
Session Chair: Tetiana Preobrazhenska (URAN)
Moderator: Irina Mikhailava (GÉANT)
Session 9 - Panel: 'Ethics in AI'
THURSDAY 26: 13:00 - 13:45