The Linguistic Blind Spot of Value-Aligned Agency, Natural and Artificial

TL;DR

This paper argues that natural language is essential for achieving robust value alignment in AI, highlighting a critical linguistic aspect often overlooked in designing ethical artificial agents.

Contribution

It introduces the idea that linguistic communication is a necessary condition for effective value alignment in AI systems, impacting future research directions.

Findings

Linguistic communication is crucial for value alignment.

Current AI alignment efforts may overlook the importance of language.

Natural language understanding is key to ethical AI development.

Abstract

The value-alignment problem for artificial intelligence (AI) asks how we can ensure that the 'values' (i.e., objective functions) of artificial systems are aligned with the values of humanity. In this paper, I argue that linguistic communication (natural language) is a necessary condition for robust value alignment. I discuss the consequences that the truth of this claim would have for research programmes that attempt to ensure value alignment for AI systems; or, more loftily, designing robustly beneficial or ethical artificial agents.