AI-as-exploration: Navigating intelligence space
This work addresses the need for a broader scientific perspective on AI's potential to reveal diverse forms of intelligence, which is foundational for AI researchers and theorists.
The paper tackles the problem of defining AI's role in exploring intelligence beyond human and animal forms, proposing 'AI-as-exploration' as a framework and illustrating it with a case study showing that Large Language Models achieve human-level accuracy in concept combination tasks but through different mechanisms.
Artificial Intelligence is a field that lives many lives, and the term has come to encompass a motley collection of scientific and commercial endeavours. In this paper, I articulate the contours of a rather neglected but central scientific role that AI has to play, which I dub `AI-as-exploration'.The basic thrust of AI-as-exploration is that of creating and studying systems that can reveal candidate building blocks of intelligence that may differ from the forms of human and animal intelligence we are familiar with. In other words, I suggest that AI is one of the best tools we have for exploring intelligence space, namely the space of possible intelligent systems. I illustrate the value of AI-as-exploration by focusing on a specific case study, i.e., recent work on the capacity to combine novel and invented concepts in humans and Large Language Models. I show that the latter, despite showing human-level accuracy in such a task, probably solve it in ways radically different, but no less relevant to intelligence research, to those hypothesised for humans.