在ＡＩ醫生和病人之間——人工智能診斷技術的內在邏輯及其對病人主體性建構的影響

LANGUAGE NOTE | Document text in Chinese; abstract also in English.近年來，人工智能（ＡＩ）技術在醫療領域中的應用探索十分活躍，ＡＩ診斷又是其中的關注焦點。它可以使診斷更加高效和準確，從而在整體上改善醫療診斷品質，提高醫療資源的供给效率和公平性。不論從ＡＩ診斷發展的技術邏輯還是西方現代醫學診斷思維來看，從智慧助手發展到ＡＩ醫生都具有一定的合理性與必然性。但這一技術具有其内在局限，即使已經發展到理想狀態，ＡＩ醫生也不能完全取代醫生，它能輔助醫生和患者更充分地發揮各自的能動性，卻有可能傷害醫患互動的關鍵内核：意義生成和人格構建。如何在充分利用ＡＩ診斷醫生的效率的同時避免其有可能帶來的負面影響，人們需要對ＡＩ診斷技術發展的底層邏輯提供足夠有力的理論拮抗，而不是停留在技術所限定好的語境中解決那些具體的問題。本文提出，有必要引入中國傳統醫學的診斷思維和儒家“成人”理論，作為針對西方現代醫學觀念和現代主體性哲學發起反思的重要理論資源，並在此基礎上重新審視技術與人文的關係。Applying AI in medical contexts, especially for diagnosis, has become very popular in recent years. AI has the potential to make diagnosis more efficient and accurate, improving the overall quality of medical diagnosis and making medical provisions fairer and more effective. Combining the logic of AI with that of modern Western medical diagnosis, it is to some extent intuitive to imagine AI physicians. However, even in its ideal form, AI technology has intrinsic limitations that will prevent it from completely replacing physicians. Although AI can help physicians and patients to develop their own agency, it may strike at the core of physician–patient interaction: generating meaning and constructing personhood/subjectivity. How can we make best use of the efficiency of AI diagnosis while avoiding its potential negative influence? There needs to be a powerful theoretical rejoinder to the fundamental logic of AI diagnosis. It is not enough to deal with specific issues within the realm already delimited by AI diagnosis technology. This paper highlights the need to incorporate the way of thinking of traditional Chinese medical diagnosis and the Confucian theory of “the way of becoming a person.” Both are important theoretical resources that can be used to counterbalance the way of thinking of modern Western medicine and modern Western philosophy, which emphasize subjectivity. On this basis, the relationship between technology and the humanities can be re-examined.DOWNLOAD HISTORY | This article has been downloaded 65 times in Digital Commons before migrating into this platform.