The Artificial Intelligence and Cognitive Science Working Group (AI CogSci WG) is part of the Technical Committee 12 Artificial Intelligence, which is one of several technical committees of the International Federation of Information Processing (IFIP).
These committees contribute to and drive progress in the state of the art in their respective area, usually by forming working groups, i.e., groups that bring together experts in the field, organize meetings, and generally contribute to pushing research onward.
Historically, cognitive science and artificial intelligence (AI) had many overlaps and developed nearly in parallel, with many key people involved in both areas and many common research interests. Consider the 1956 workshop on Artificial Intelligence at Dartmouth College (often considered the seminal event of AI as a research field) and the Symposium on Information Theory at MIT (often seen as `the birth’ of cognitive science as a field) in the same year, with many overlapping participants.
While these strong links still exist in terms of common research problems and interests, the `visible’ links seem to have become weaker in recent years, i.e., having strong representations of both disciplines at the same conference or joint research projects.
However, we argue that in today’s age of ‘black-box’ AI, implemented in self-driving vehicles, service and social robots, interactive location based services, and smart cities (among others), a strong connection between Cognitive Science and AI researchers is more important than ever. Most of these systems, as well as other future AI systems, will perform tasks while directly immersed in our day-to-day environments; many of these tasks will also be predominantly spatial or will have strong spatial components.
People will likely perceive these systems as social entities, and will expect to interact with them in predicable, ‘rational’ ways. Thus, the systems will need to understand relevant human concepts about communication in general, and about space in particular, and should be able to explain their behavior in ways people are familiar with. This calls for explainable AI, but also for understanding principles of human reasoning, representation, and communication. In other words, we strongly believe it needs a concerted effort of both AI and Cognitive Science researchers to develop truly autonomous, smart, socially accepted AI systems that will function in our unstructured, ‘messy’ everyday world.