The paper is dedicated to the study of the theoretical and technological development that occurred, in particular in the XX century, in the sector of Artificial Intelligence. According to the theoretical framework of mechanical rationalism, we study how the development of machine intelligence is a continuation, through different means, of the old process of outsourcing of cognitive activities by humans onto parts of their physical environments. Because of this process, an increasingly larger portion of the non-human environment performs perceptive and cognitive activities. From this follows that machine systems, not necessarily humans anymore, are the components of the physical environment that perform measurements on the universe of which the humans are also components. We suggest that the scientific discussion on the topic of AI development could be framed in the context of a more general phenomenon of an increase in the computational and perceptual capabilities of the physical universe, as opposed to a merely human and technological problem. This is because, ever so slightly, humans are being removed from the cognitive processes of technological systems they created, which continue to perceive and think autonomously. The act of machine cognition, or rather, of machine measurements, causes an effect on the environment in which humans live, and ever more so than the human measurements. Finally, we discuss the current approach to the development of viable AI systems that aim at increasing the reciprocal intelligence of humans and machines, rather than the replacement of the former's cognitive faculties by the latter.

本文致力于研究人工智能领域的理论和技术发展，特别是在20世纪。根据机械理性主义的理论框架，我们研究了机器智能的发展是如何通过不同的方式延续人类将认知活动外包到其部分物理环境的旧过程。由于此过程，越来越多的非人类环境执行感知和认知活动。由此可见，机器系统（不再是人类）已成为物理环境的组成部分，它们在人类也是其组成部分的宇宙上执行测量。我们建议，关于人工智能发展主题的科学讨论可以在一个更普遍的现象中进行，即物理宇宙的计算和感知能力增加，而不是人类和技术问题。这是因为，人类被创造出的技术系统的认知过程被移出了人类的技术过程，这种过程持续不断地自主地感知和思考。机器认知的行为，或者说是机器测量的行为，对人类生活环境的影响比人类测量的影响更大。最后，我们讨论了可行的AI系统开发的当前方法，该方法旨在提高人与机器的交互智能，而不是替代前者。