Constanza Fierro

AI-generated text, images, videos, and music are a growing presence on our cultural landscape. But are AI models ever artists, that is, the authors of works of art? While many philosophers allow that human artists can use generative AI systems to produce artworks, almost all reject the possibility of current AI models authoring those works themselves. A central argument for this conclusion invokes intentions: authoring a work of art is an essentially intentional process, and current AI models lack intentions; thus, current AI models do not author artworks. In this paper, we scrutinize the premises of this argument. We distinguish three kinds of intentions that have been proposed as necessary for artistic authorship. We highlight empirical evidence suggesting that, contrary to prevailing assumptions, all three kinds of intentions may arise in at least some generative AI models. We thus conclude that, if there remains a significant barrier to AI models’ having the capacity to author artworks, it is unlikely to turn on their lack of intentions.

Mechanistic interpretability (MI) aims to explain how neural networks work by uncovering their underlying mechanisms. As the field grows in influence, it is increasingly important to examine not just models themselves, but the assumptions, concepts and explanatory strategies implicit in MI research. We argue that mechanistic interpretability needs philosophy as an ongoing partner in clarifying its concepts, refining its methods, and navigating the epistemic and ethical complexities of interpreting AI systems. There is significant unrealised potential for progress in MI to be gained through deeper engagement with philosophers and philosophical frameworks. Taking three open problems from the MI literature as examples, this paper illustrates the value philosophy can add to MI research, and outlines a path toward deeper interdisciplinary dialogue.

Knowledge claims are abundant in the literature on large language models (LLMs); but can we say that GPT-4 truly "knows" the Earth is round? To address this question, we review standard definitions of knowledge in epistemology and we formalize interpretations applicable to LLMs. In doing so, we identify inconsistencies and gaps in how current NLP research conceptualizes knowledge with respect to epistemological frameworks. Additionally, we conduct a survey of 100 professional philosophers and computer scientists to compare their preferences in knowledge definitions and their views on whether LLMs can really be said to know. Finally, we suggest evaluation protocols for testing knowledge in accordance to the most relevant definitions.