Frontiers | Artificial Intelligence Is Stupid and Causal Reasoning Will Not Fix It

Artificial Neural Networks have reached ‘Grandmaster’ and even ‘super-human’ performance’ across a variety of games, from those involving perfect-information, such as Go ((Silver et al. (2016)); to those involving imperfect-information, such as ‘Starcraft’ (Vinyals et al. (2019)). Such technological developments from AI-labs have ushered concomitant applications across the world of business, where an ‘AI’ brand-tag is fast becoming ubiquitous. A corollary of such widespread commercial deployment is that when AI gets things wrong - an autonomous vehicle crashes; a chatbot exhibits ‘racist’ behaviour; automated credit-scoring processes ‘discriminate’ on gender etc. - there are often significant financial, legal and brand consequences, and the incident becomes major news. As Judea Pearl sees it, the underlying reason for such mistakes is that “... all the impressive achievements of deep learning amount to just curve fitting”. The key, Pearl suggests (Pearl and Mackenzie (2018)), is to replace ‘reasoning by association’ with ‘causal reasoning’ - the ability to infer causes from observed phenomena. It is a point that was echoed by Gary Marcus and Ernest Davis in a recent piece for the New York Times (Marcus and Davis (2019)): “we need to stop building computer systems that merely get better and better at detecting statistical patterns in data sets – often using an approach known as “Deep Learning” – and start building computer systems that from the moment of their assembly innately grasp three basic concepts: time, space and causality”. In this paper, foregrounding what in 1949 Gilbert Ryle termed ‘a category mistake’ (Ryle (1949), pp. 16), I will offer an alternative explanation for AI errors; it is not so much that AI machinery cannot ‘grasp’ causality, but that AI machinery (qua computation) cannot understand anything at all.

Related Keywords

New York , United States , Japan , Erhan , Jiangsu , China , United Kingdom , Andover , Hampshire , Washington , Stonham , Suffolk , Melbourne , Victoria , Australia , Irvine , California , London , City Of , San Francisco , Nob Hill , Berkeley , Stockholm , Sweden , Hollywood , Vienna , Wien , Austria , Shelter Island , Chinese , Japanese , American , Jaquet Droz , Warren Mcculloch , Stephen Hawking , John Burgess , Hilary Putnam , David Chalmers , Francis Crick , Ray Kurzweil , Kevin Warwick , Paul Benacerraf , Peter Norvig , Alan Turing , Jm Bishop Oxford , Francois Chollet , Yoshua Bengio , Georges Rey , Xiaoice Chatbot , S Simpson Cambridge , Judea Pearl , John Lucas , Roger Penrose , Stefan Harnad , Selmer Bringsjord , Tomaso Poggio , John Searle , Mcculloch Pitts , Cl Nehani Heidelberg , Lewis Carroll Alice , Walter Pitts Mcculloch , Yann Lecun , American Philosophical Association , Google , Artificial Neural Networks London , Association For Computing Machinery , International Conference On Learning Representations , Microsoft Research , Cambridge University Press , International Conference On Machine , Nob Hill Masonic Center , Variational Autoencoder Networks , Bayesian Networks , Artificial Neural Network , Adaptive Networks , Cambridge University , Autoencoder Networks , Autoencoder Networks Kramer , Bayesian Networks Pearl , Variational Autoencoder Networks Kingma , Conference Of The Cognitive Science Society Irvine , University Of London , Manning Publications Co , Harvard University Press , Generative Adversarial Networks Goodfellow , University Of Vienna , Neural Networks , Basic Books , Twitter , Variational Autoencoders To Generative Adversarial Networks , Microsoft , Harvard University , Princeton University Press , Generative Adversarial Networks , Oxford University , Problem Solving Using Artificial Neural Networks , Convolutional Networks , Random House , Massachusetts Institute Of Technology , Princeton University , Autoencoder Network , Generative Autoencoder Network , Youtube , Oxford University Press , Warner Brothers , Springer International Publishing , Artificial Neural Networks , Routledge , Lewis Carroll , Turing Machine , Multi Layer Perceptrons , Walter Pitts , Layer Perceptrons , Variational Autoencoders , Generative Adversarial , Terence Broad , Blade Runner , Ridley Scott , Generative Architectures , Adversarial Networks , Solving Using Artificial Neural , Artificial General Intelligence , Massachusetts Institute , Singularity Summit , Deep Learning , Deep Neural Networks , Microsoft Xiaoice , Personal Assistants , Apple Siri , Amazon Alexa , Google Assistant , Take Xiaoice , Core Chat , Talk About , Peter Lee , Corporate Vice President , Microsoft Healthcare , Human Knowledge , Chinese Room , Sir Roger Penrose , New Mind , Basic Penrose Argument , Del Sentence , Professor Kevin Warwick , Dancing With Pixies , Jaquet Droz The Writer , Computing Machinery , State Machine , Finite State Automaton , Jaquet Droz Writer , Finite State , Cengage Learning , Automata Studies , International Conference , Machine Learning , Med Abstract , Crossref Full Text , New Essays , Artificial Intelligence , Computational Intelligence , Life Support Systems , Eolss Publishers , Journeys Beyond , Turing Barrier , Eastern Joint Computer Conference , Annual Workshop , Computational Learning Theory , Radial Basis Functions , Multi Variable Functional Interpolation , Radar Establishment , His Centennial , Manning Publications , Astonishing Hypothesis , Scientific Search , Causal Reasoning , Neural Information Processing Systems , Times Literary Supplement , Learning Representations , Computational Mind , Feminist Uncanny , Bloomsbury Studies , Bloomsbury Academic , Singularity Is Near , Humans Transcend , Nob Hill Masonic , Computer Supported Cooperative Work , Social Computing , Cognitive Science Society , New Science , Concerning Computers , Missing Science , Enrose Lucas Argument , Ausal Cognition , Rtificial Neural Networks , Cognitive Science , Chinese Room Argument ,