杭士基與經濟學人說AI

Chomsky’s and the Economist’s discourse on the Artificial Intelligence

開創生成語言學的杭士基是AI先驅研究者，秉持自由左翼的世界觀；《經濟學人》週刊自稱激進中間派。本文分析ChatGPT發表後兩整年，周刊共152篇AI文章，以及，杭士基及其評論者在此後與先前的AI言說與文字。作者論證，雙方相同的部分至少有五：生成語言學未因AI而過時、人與AI產品都存在幻覺與黑洞、作為AI重要產品的LLM在特定領域用途明顯、一刊一人都寡談AI的烏托邦或反烏托邦而是注重眼前大事，以及，AI加重現存隱私與著作權……等等難題。雙方對AI是否能夠正德厚生的認知，分野有二。首先，杭士基強調而刊物較不凸顯的是，AI若有效益，社會能否因而雨露均霑，關鍵不在技術，而是決斷於創造與使用AI產品的倫理與社會體制。其次，與前相關，周刊關注AI是否能夠提高生產力，並且據此問題意識而分析過往對新技術與當前對AI投資過熱的表現，見解值得重視。杭士基在2023年中風，本文取用杭士基觀察與評論人文社會現象六十年當中，可與刊物AI觀對話的部分。準此，本文論稱，周刊儼然認定而杭士基不會同意的看法有二：富國的生產力提高仍然是氣候變遷與地球環境生態可以承擔；AI等技術即便造成創造性毀滅，人類的經濟體系最終仍會回歸平衡。

Noam Chomsky, a pioneer in generative linguistics and an early AI researcher, holds a liberal-left worldview. This article analyzes 152 AI-related pieces published over two full years by The Economist, a self-proclaimed radical centrist publication, in the wake of ChatGPT’s release. It also examines Chomsky’s own writings and those of his commentators—both past and present—on AI, comparing the similarities and differences between the two perspectives.
The author argues that there are at least five key areas of convergence: Generative linguistics remains relevant despite the rise of AI. Both humans and AI products are prone to hallucinations and blind spots.
Large Language Models (LLMs), as major AI products, show clear utility in specific domains. Rather than indulging in utopian or dystopian visions of AI, both focus on pressing contemporary issues. AI exacerbates existing challenges around difficult issues such as privacy and intellectual property…...etc. However, their views diverge in two significant ways regarding whether AI can contribute to moral and social flourishing. First, Chomsky emphasizes—more than the publication does—that if AI is to be beneficial, the key lies not in the technology itself but in socialist ethics and institutions. Second, The Economist focuses on whether AI can enhance productivity, and analyzes historical responses to new technologies and the current overheated investments surrounding AI. In comparison, the article draws on Chomsky’s observations and commentaries on human and social phenomena from the past six decades to create a dialogue. Based on this, the article contends that while the Economist seems to take it for granted, Chomsky would not agree that the productivity gains of wealthy nations are something that climate change and the global environment can afford. Furthermore, he would not agree that even if AI and other technologies cause creative destruction, the human economic system will ultimately return to equilibrium.