Dr LaCroix's article is published by the Philosophical Studies Journal
In this (open access) paper, LaCroix argues that language is necessary for robust value alignment, exploring the consequences of this claim in the specific context of the value alignment problem for artificial intelligence.
The value alignment problem for artificial intelligence asks how we can ensure that the "values"—i.e., objective functions—of artificial systems are aligned with the values of humanity. In this (open access) paper, LaCroix argues that linguistic communication is a necessary condition for robust value alignment. This research is related to LaCroix's book on Artificial Intelligence and the Value Alignment Problem, which will be published in Spring 2025.
The article further discusses 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, those programmes that seek to design robustly beneficial or ethical artificial agents. Minimally, the necessity of linguistic communication for robust value alignment would set a significantly burdensome lower bound on the very possibility of aligning values.
The article is published in Philosophical Studies as part of their collection on Normative Theory and Artificial Intelligence.
Find out more
- Learn more about Dr Travis LaCroix
- Read the full article published in Philosophical Studies Journal.
- Our Department of Philosophy is ranked 29th in the QS World University Rankings by Subject 2024. Visit our Philosophy webpages for more information on our undergraduate and postgraduate programmes.