Intelligence without Robots: A Reply to Brooks

AI: “The agents should be embodied as mobile robots...the new approach can be extended to cover the whole story, both with regards to building intelligent systems and to understanding human intelligence” (Brooks 1991a, p. 585). This article argues that even if we accept Brooks’s first position and seek to build complete agents in realworld environments, we need not accept robotics as the foundation for AI. Clearly, robotics is an important enterprise, with much to contribute to AI. However, I challenge Brooks’s position that the primary path to progress in AI is “to study intelligence from the bottom up, concentrating on physical systems (e.g., mobile robots), situated in the world, autonomously carrying out tasks of various sorts” (Brooks 1991a, p. 569). In this article, I propose real-world ■ In his recent papers, entitled “Intelligence without Representation” and “Intelligence without Reason,” Brooks argues for mobile robots as the foundation of AI research. This article argues that even if we seek to investigate complete agents in real-world environments, robotics is neither necessary nor sufficient as a basis for AI research. The article proposes real-world software environments, such as operating systems or databases, as a complementary substrate for intelligent-agent research and considers the relative advantages of software environments as test beds for AI. First, the cost, effort, and expertise necessary to develop and systematically experiment with software artifacts are relatively low. Second, software environments circumvent many thorny but peripheral research issues that are inescapable in physical environments. Brooks’s mobile robots tug AI toward a bottom-up focus in which the mechanics of perception and mobility mingle inextricably with or even supersede core AI research. In contrast, the softbots (software robots) I advocate facilitate the study of classical AI problems in real-world (albeit, software) domains. For example, the UNIX softbot under development at the University of Washington has led us to investigate planning with incomplete information, interleaving planning and execution, and a host of related high-level issues.