A Plea for Inexplicability
DOI:
https://doi.org/10.6092/issn.1971-8853/15296Keywords:
AI, Interpretability, Explicability, Philosophy, ChaosAbstract
Explanations have long played a crucial role in the west’s metaphysics of control by providing levers by which we exercise control, but also by grounding our assumption that we are the legitimate and rightful masters of the world. Within this metaphysics, inexplicability looks like a failure. But machine learning may be teaching us a different lesson: Explicability is not a property of the universe, but inexplicability is. In short, the world is the ultimate black box. Accepting this may lessen our one-sided commitment to the general rules, principles, and laws that make western-style mastery seem possible. We may instead be entering a time when we are willing to value the particular at least as much as the general, and can embrace the hiddenness of the universe that grounds all that shows itself to us.
References
Doshi-Velez, F., & Kim, B. (2017). Towards a Rigorous Science of Interpretable Machine Learning. arXiv, 2 March. https://arxiv.org/pdf/1702.08608.pdf
Douglas, H.E. (2009). Reintroducing Prediction to Explanation. Philosophy of Science, 76(4), 444–463. https://doi.org/10.1086/648111
Fell, J. P. (1979). Heidegger and Sartre: An Essay on Being and Place. New York, NY: Columbia University Press. https://doi.org/10.7312/fell91382
Heidegger, M. (2006). Sein und Zeit (19th ed.). Tübingen: Niemeyer. (Original work published 1927)
Heidegger, M. (2012). Der Ursprung des Kunstwerkes. Frankfurt am Main: Klostermann. (Original work published 1935-6)
Linardatos, P., Papastefanopoulos, V., & Kotsiantis, S. (2021). Explainable AI: A Review of Machine Learning Interpretability Methods. Entropy, 23(1), 1–45. https://doi.org/10.3390/e23010018
Mittelstadt, B., Wachter, S., Russell, C., & Sutcliffe, D. (2018). Could Counterfactuals Explain Algorithmic Decisions without Opening the Black Box? Oxford Internet Institute, 15 January. https://www.oii.ox.ac.uk/news-events/news/could-counterfactuals-explain-algorithmic-decisions-without-opening-the-black-box/
Weinberger, D. (1984). Earth, World and Fourfold. Tulane Studies in Philosophy, 32, 103–109. https://doi.org/10.5840/tulane19843210
Published
How to Cite
Issue
Section
License
Copyright (c) 2022 David Weinberger
This work is licensed under a Creative Commons Attribution 4.0 International License.