如何成為「人」：缺陷及其經驗作為對人工智能研究之啟發－以自動駕駛技術為例

＂Becoming Human＂: the idea of fault and its experiences as an inspiration for artificial intelligence - cases from self-driving cars

劉育成(Yu-cheng Liu)

《資訊社會研究》 38期 (2020/01) Pp. 93-125

https://doi.org/10.29843/JCCIS.202001_(38).0006

摘要

本研究以技術人類學家Bernard Stiegler有關人類因Epimetheus的「失誤」而得以「成為人」的論述，對人類智能（human intelligence）與人工智能（artificial intelligence）進行探討。以「失誤」概念為基礎的「缺陷」及對其之經驗，或許是「人之所以為人」的關鍵元素。人類不僅藉由技術的外部化，以彌補其「失誤」，更透過技術來區別自身於其他物種，進而肯定「成為人」的意義。本研究提出，若將Stiegler的雙關「去－失誤／預設」（de-fault／default）視為一組區別，那麼從人類智能到人工智能的發展，即是一個從「de-fault／default」到「default／de-fault」的轉變，該轉變有助於理解人工智能現象的內涵。本研究以自動駕駛技術為例，探討其困境與未來發展。在代結論中，我將提出以「例示／窮盡」（illustration／exhaustion）之區別，作為對人工智能現象的反思。

關鍵字

人類智能 ； 人工智能 ； 技術人類學 ； 去失誤／預設 ； 自動駕駛技術

並列摘要

This paper examines the ideas of fault and de-fault developed by Bernard Stiegler in his discussion of the relationship between human beings and technology. ＂Becoming human＂ means how we as human beings encounter and cope with what Stiegler terms ＂original default＂. In his discussion, human beings appear in the externalization of this original de-fault, and in doing so human beings can survive with various technologies which are products of this externalization. It becomes the original de-fault for human beings and later ＂default＂. How to become humans or how machines can think like humans refers to the transformation from ＂de-fault/default＂ to ＂default/de-fault＂ in terms of the theory of form. With the perspective from the anthropology of technology, mainly extended from the idea of Bernard Stiegler, and with the examination of the technologies of artificial intelligence applied around self-driving cars, the argument of this paper may contribute to the future thinking of artificial intelligence. In addition, as one of concluding remarks, the author also suggests another distinction of illustration and exhaustion as a possible way to think about artificial intelligence in the future.

並列關鍵字

human intelligence ； artificial intelligence ； the anthropology of technology ； de-fault/default ； driverless technology

參考文獻

Bernecker, S., & Michaelian, K. (Eds.). (2017). The Routledge Handbook of Philosophy of Memory. London and New York: Routledge. https://doi.org/10.4324/9781315687315

Bostrom, N. (2014). Superintelligence: Paths, Dangers, Strategies. Oxford, UK: Oxford University Press. DOI: https://doi.org/10.1017/S0031819115000340

Clark, A. (2013). Re-Inventing Ourselves: The Plasticity of Embodiment, Sensing, and Mind. In M. More & N. Vita-More (Eds.), The Transhumanist Reader. West Sussex, UK: Wiley-Blackwell. https://doi.org/10.1002/9781118555927.ch11

Dreyfus, H. L. (2006). Overcoming the myth of the mental. Topoi, 25, 43-49. https://doi.org/10.1007/s11245-006-0006-1

Dreyfus, H. L. (2007). Why Heideggerian AI failed and how fixing it would require making it more Heideggerian. Artificial Intelligence, 171, 1137-1160. https://doi.org/10.1016/j.artint.2007.10.012