@InProceedings{Sloman1995,
title = "A Philosophical Encounter",
author = "Aaron Sloman",
year = "1995",
bibdate = "2002-01-03",
bibsource = "DBLP,
http://dblp.uni-trier.de/db/conf/ijcai/ijcai95.html#Sloman95",
booktitle = "IJCAI",
pages = "2037--2040",
}
Philosophy needs AI in order to make progress with many difficult questions about the nature of the mind, and AI needs philosophy in order to help clarify goals, methods, and concepts and to help with several specific technical problems.
AI as philosophyMost AI researchers regard philosophy as irrelevant to their work, while others think of the two as very closely related
Both philosphy and AI asks the questions: what sort of minds are possible? and what makes them possible in a physical world?
'Mind' is a cluster concept referring to an ill defined collection of features, rather than a single property that is either present or absent. Since different collections of features define different minds, the old philosohpical task of explaining what a mind is, is replaced by exploration of what minds are by studying their mechanisms, capabilities, how they develop, and how some might evolve.
AI can advance philosophy also in the area of "qualia", also referred to as "raw feels". These are the contents of our experiences, the answer to what it is like to feel, see or want something, and so on.From an AI perspective a completely autonomous agent must obviously have information stores, both long term collections of general information, and short term stores of current goals, plans, thoughts, etc. It is less obvious that a completely autonomous agent needs to be able to attend to and control some of its internal databases, and may need to inform others about them. Attention to internal states may play an important role in learning. Philosophers who refer to 'qualia' are referring to these internal states.
In philosophy there are unanswered questions about the nature of mathematical concepts and knowledge. AI can gain understanding by doing experimental epistemology and studying design for human-like information processing architectures that cna learn about numbers in the way that children do. In other words, AI by studying how these mathematical concepts are learned by intelligent agents can give understanding to these age old questions of philosophy.
A new potentially important area of influence of AI on both philosophy and psychology concerns the study of motivation and emotions. As AI researchers design agents capable of motivation and emotions they will have a better understanding of how these things arise in the mind.
Philosophy as AI:Philosophy needs AI to help with age-old problems, but AI needs philosophy also. Ai already builds on work by philosophers. Some ways AI uses philosophy: the use of speech act theory, Kant's proof of in Critique of Pure Reason. Another development is the recognition of deep connections between the AI task of understanding what sort of knowledge an intelligent system requires and the older philosophical activities of metaphysics.
Two way influences and more:In all the above listed areas AI and philosophy inform each other, and as the depth of analysis increases, the amount of feedback increases, the work becomes more technical and specialised and the boundary between the two will disappear.
Insights from AI will eventually make a deep impact on psychotherapy, because of AI's role in rationality in intelligence.
There is a disappearing boundary between AI and philosophy. Often boundaries between diciplines exist in the academic world just because of politics, or the organisation of research funding agencies - not because there is a clear cut boundary that exists. The topics discussed in this article will need to merge further, and with other disciplines such as neurpscience, psychology, and social science.
Other areas in which philosophy and AI overlap: consciousness and free will, how machines understand the symbols they are given, the relevance of mathematical incompleteness theorems, confusions surrounding the Turing test, the role of states like pain and pleasure in intelligent agents, ehtical issues about the rights and responsibilities of intelligent artefacats, debates about the philosophical significance of the choice between connectionist implementations and symbolic implementations, whether mentality requires causal embedding in an external physical environment, whether AI needs non-computational as well as computational mechanisms, analysis of the concept of 'computation', and prospects for future forms of intelligence.