“Can Machines Think?” A Cautionary View on A(n)I (Ethics Seminar)

Abstract:

Alan Turing’s paper ‘Computing Machinery and Intelligence’ (Mind 49: 433-460, 1950) starts with the famous question ‘Can machines think?’ However, immediately after the question is stated, Turing warns that using the common meanings of the words ‘machine’ and ‘think’ is dangerous and thus he replaces the question with the imitation game in which a human must decide, based only on written conversation, whether their interlocutor is a computer or another human.

The imaginary game is built by ‘drawing a fairly sharp line between the physical and the intellectual capacities of a man’ which means that the human cannot rely on their full inherent human attributes usually described by words like ‘think’, ‘learn’, and ‘intelligence’. The dramatic restrictions imposed on the human player provided practical pathways not only for building computers but also for programming the so-called ‘artificial intelligence’ (AI) algorithms.

Turing’s belief that by the end of the twentieth century ‘the use of words […] will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted’ combined with apparent language ambiguities lead to today’s misconception that AI models could become more intelligent than humans. The aim of this talk is twofold. First, an attempt will be made to clarify the language used to describe human and computer attributes by providing a very brief presentation of how AI models work and current definitions of some words expressing continuously evolving human qualities and concepts. Second, certain worrying aspects of AI development and usage and their dire consequences to humanity will be mentioned.

Bio:

Dr. Drapaca is an Associate Professor of Engineering Science and Mechanics at Penn State University. She has B.S. and M.S. degrees in Applied Mathematics from University of Bucharest, Romania, and a Ph.D. degree in Applied Mathematics from University of Waterloo, Canada. Before joining Penn State, Dr. Drapaca held postdoctoral fellowships in the Department of Radiology at University of California in San Francisco and in the Department of Physiology and Biomedical Engineering at Mayo Clinic.

Her expertise is in mathematical modeling in medicine, continuum and non-local mechanics, medical image analysis, and inverse problems. The focus of her research has been understanding mechanisms of onset and evolution of brain diseases through multiphysics and multiscale mathematical models and corresponding numerical simulations. She is the coauthor of a Fields Institute monograph on mathematical modelling and biomechanics of the brain.