Is absence of affect a barrier to AGI and alignment?

Question

The question discusses Damasio's work on affective neuroscience and its implications for AGI and alignment, wondering if the absence of affect in current AI systems is a fundamental barrier to achieving true general intelligence and alignment.

HackerNews · Accepted Answer

The absence of affect in current AI systems may indeed be a significant barrier to achieving true general intelligence and alignment. Damasio's work on somatic markers highlights the crucial role of emotions in decision-making, suggesting that intelligence is not just about computation, but also about the ability to assign significance to different outcomes. Current AI systems, such as LLMs, lack this felt sense of significance, which may limit their ability to make decisions and be truly aligned with human values.

Is absence of affect a barrier to AGI and alignment?

Resources