City Research Online

Abstract

In this chapter the notion of agency in AI is presented..It has been argued that in order to behave rationally in prevalent software applications artificial entities would have to be autonomous and adaptive. Besides, rather than working with single, isolated systems the new trend in AI would need to focus on inherently social entities in the form of multi-agent systems.

The chapter begins by introducing the notion of action in traditional AI systems, deliberative and reactive. Next, the idea of agency is presented as complementary to classical AI approaches to action as well as a key factor in the development of applications and services that currently rely on technologies where software systems execute instructions automatically. In particular, the chapter highlights the importance of developing theories of action and learning in multi-agent scenarios such as the Internet. This introduction shall conclude with some considerations about the research areas that need to be targeted if the agent paradigm is to become the standard in the design, specification and implementation of intelligent systems.

Publication Type:	Book Section
Additional Information:	Copyright Cambridge University Press, 2014. Reprinted with permission.
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	School of Science & Technology > Computer Science

Export

Downloads