Autonomous vehicles (AVs) have received a great deal of attention in recent years, with many commentators asking how these vehicles will “see” to navigate themselves and, more important, avoid colliding with people, other vehicles, and objects. This article analyzes how AVs see and the data sets they create through Jean Baudrillard’s framework of simulation and simulacra, paying attention to how these technologies purify and classify the real into a series of representations. In doing so, it extends the existing critique of data-generating “smart” technologies and automation and additionally draws on visual theory to understand how the hyperreality of AVs is constructed. Technologies employed on AV trials such as light detection and ranging videos and virtual reality produce data sets that create a version of the urban environment that then become the model for decisions regarding the operations and management of the city. Yet, as Baudrillard theorized, such models risk being mistaken for the reality that they represent and the data set itself confusing the nature–culture divide and created hybrid geographies of the city. As a simulacrum that operates on its own underlying logic, the article argues that this version of seeing or visioning the world risks rendering it as an object of scrutiny for state and nonstate actors that override the realities and people of the city. Indeed, techno-visions and testing of AVs can lead to situations in which reality itself is manipulated to conforming with the simulation, further complicating the simulacra of the urban environments in which AV testing takes place.

近年来，自动驾驶汽车 (AV) 受到了极大关注，许多评论员询问这些车辆将如何“看到”以进行自我导航，更重要的是，如何避免与人、其他车辆和物体发生碰撞。本文通过让·鲍德里亚 (Jean Baudrillard) 的模拟和拟像框架分析了 AV 是如何看到的以及它们创建的数据集，关注这些技术如何将真实净化和分类为一系列表示。在这样做的过程中，它扩展了对数据生成“智能”技术和自动化的现有批评，并另外利用视觉理论来理解 AV 的超现实是如何构建的。AV 试验中采用的技术，例如光检测和测距视频以及虚拟现实，产生的数据集可以创建城市环境的一个版本，然后成为有关城市运营和管理决策的模型。然而，正如鲍德里亚的理论，这些模型有可能被误认为它们所代表的现实和数据集本身混淆了自然-文化鸿沟并创造了城市的混合地理。作为一种根据其自身基本逻辑运作的拟像，文章认为，这种看待或想象世界的版本有可能使其成为国家和非国家行为者的审查对象，这些行为者凌驾于城市的现实和人民之上。事实上，自动驾驶汽车的技术愿景和测试会导致现实本身被操纵以符合模拟的情况，