社会和技术属性与算法偏向

Social and Technical Properties and Algorithmic Bias

算法的技术性沿袭并外化了算法的生物性表征,建立了诸多算法物质性、社会性层面的关系;算法的社会性,是算法依托技术逻辑而形成的制度延展或文化实践。算法不只是严格限制和程序公式,不是超越人类网络的行为体,不是自主的技术对象,而是复杂的社会技术系统。人类所具有的善恶、文化差异、意识偏见,也会反映于算法之中。算法的偏向,除了技术的偏向、工程师可能的人为偏向外,最大的偏向往往出自于技术与社会和人的互动之中。

The technicality of algorithms follows and externalizes the biological representation of algorithms, establishing many relationships on the material and social levels of algorithms;the sociality of algorithms is the institutional extension or cultural practice of algorithms based on technical logic. Algorithms are not just strictly limited and procedural formulas, not actors in a network beyond humans, not autonomous technical objects, but complex socio-technical systems. The good and evil, cultural differences, and conscious biases that humans possess are also reflected in algorithms. In addition to the technical bias and the possible artificial bias of engineers, the biggest bias of algorithms often comes from the bias formed in the interaction of technology with society and human beings.