ESHS/HSS Annual Meeting

Artificial Intelligence Philosophy: From Kenneth Sayre's Mental Simulation To Leo Breiman's Algorithmic Modeling

• In Event: Rewiring Humans 2: Biology, Psychotherapy, and Philosophy in the Age of Digital Computers

English Abstract

The emergence of the field of artificial intelligence was based on the “conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”.
The philosopher Kenneth M. Sayre, who was working on his doctoral thesis in philosophy at Harvard University while also working as a systems analyst for Oliver Selfridge and Marvin Minsky at MIT's Lincoln Laboratory, argued in his book on “Recognition” that our understanding of a particular type of human behaviour and our ability to simulate it go hand in hand. A good way to study natural intelligence, he argued, was therefore to attempt to reproduce it artificially, and this was the motivation for his own research into automatic handwriting recognition systems. Sayre stated that conceptual ambiguities of a methodological nature were blocking current simulation attempts, with the result that no improvement in these systems was in sight. He concluded that the nature of the human behaviour we are trying to simulate is also unclear to us, and we cannot have high hopes that our attempts to simulate human behaviour will be successful.
With the advent of data science and generative models, the approach of modelling artificial intelligence as a reproduction of human intelligence has been abandoned: assuming that data is generated in some way, attempts are made to simulate these abstract processes. The data generated by the simulation should resemble the data generated by the real process.

Author

• Rudolf Seising, Deutsches Museum