The Goofus & Gallant Story Corpus for Practical Value Alignment

TL;DR

This paper introduces a multi-modal dataset combining images and descriptions to teach AI systems social norms and principles, aiming to improve value alignment and prevent harm.

Contribution

It presents a novel dataset with real-life scenarios in natural language and images, designed for training socially normative AI agents.

Findings

Dataset includes curated images and descriptions of normative behavior.

Designed for training AI to understand social principles.

Aims to improve AI value alignment and safety.

Abstract

Values or principles are key elements of human society that influence people to behave and function according to an accepted standard set of social rules to maintain social order. As AI systems are becoming ubiquitous in human society, it is a major concern that they could violate these norms or values and potentially cause harm. Thus, to prevent intentional or unintentional harm, AI systems are expected to take actions that align with these principles. Training systems to exhibit this type of behavior is difficult and often requires a specialized dataset. This work presents a multi-modal dataset illustrating normative and non-normative behavior in real-life situations described through natural language and artistic images. This training set contains curated sets of images that are designed to teach young children about social principles. We argue that this is an ideal dataset to use for training socially normative agents given this fact.