從人工智慧到人造社會性：一個新社會分析學概觀

From Artificial Intelligence to Artificial Sociality: Outlooks for a New Social Analytics

本文探討人工智慧（AI）在不同領域的迅速發展及其文獻。儘管這些資料散佈於計算機科學及工程學的專業文獻中，並且在難度與應用上有很大差異，社會科學在研究AI進展方面仍然落後於計算機科學家。近年來，研究社會的學者開始關注AI在社會生活中的廣泛滲透。本文透過引入新的觀點與概念來進一步討論，首先解釋AI在人類心中呈現為三P的組合：現象、問題與階段，接著介紹對AI的定義，並闡述這個定義下的五個基本特徵。在簡短的文獻概述之後，本文發展了關於人造社會性（AS）、新社會分析學與探索AI的方法論之跨學科性的主要論點。本文的主要提問是「AI為社會帶來了什麼？」本文回顧AI與線上文化之間的關聯，指出短期內AI、AS與線上文化不會改變人類本性，但會改變人們的組織方式與社會結構。本文認為，學者的任務是學習如何導向AI以及AS以增強人類，並且在社會結構中增強人文，並社會化地組織不可避免的人機相互依存。文中也討論AI發展期間的價值問題，以及資本主義的「附帶損害」。最後，作者強調人類不是電腦，電腦也不是人類。迄今對AI的主要缺陷是科學家尋找電腦與人類行為之間的相似之處，但實際需求恰恰相反：科學必須定義人類與機器的差異、它們各自的獨特性，以及如何透過機器與AI豐富人性。

There is an exponentially growing literature on AI that crosscuts several different fields. However, much of this material is widely scattered throughout professional computer sciences and engineering literature and varies considerably in terms of level of difficulty and substantive application. Social sciences are way behind computer scientists in surveying the reality of AI progress. However, in recent years, scholars who study society have started to look at the widespread penetration of AI instruments in the reality of societal life. This paper develops further the discussion by introducing new visions and conceptual perceptions. It starts with an explanation that AI appears in the mind of human beings as a combination of the three Ps: Phenomenon, Problem, and Phrase. The author then introduces his definition of AI and formulates five fundamental characteristics this definition underlines. After a short literature overview, the author develops his main arguments about artificial sociality (AS), new social analytics, and methodological cross-disciplinarity (potentially a-disciplinarity) in exploring AI. In what follows, the principal question is“What does AI bring to society ?”The author reviews the correlation between AI and online culture. He points out that in the short term, AI, AS, and online culture will not change human nature, however, they will change the way people are organized and society is structured. The author argues that the task for scholars is to learn how to orient AI and AS to augment humans, enhance humane in social structures, and socially organize inevitable human-machine interdependence. The value alignment problem and‘collateral damage of capitalism’during the time of AI advancement are discussed in this regard. In conclusion, the author emphasizes that Humans are not Computers, and Computers are not Humans. The primary deficiency of hitherto-existing views on AI is that scientists look for similarities between what and how computers and humans are doing. However, the real need is the opposite: science must define what are the differences between humans and machines, what makes them unique, and how to enrich humanity via machines and AI.