AI is still comparatively young, and like every young interdisciplinary area, it evolves as new ideas and insights are revealed and new methods and techniques become available. IHMC researchers James Allen, Kenneth Ford, Clark Glymour, Patrick Hayes, Choh Man Teng, Brent Venable, and Yorick Wilks actively are interested in clarifying the philosophical foundations of AI and cognitive science.

Rethinking the aims of AI

Traditionally, the goal of artificial intelligence was the creation of a ‘human-level’ mechanical intelligence, a kind of simulated human being that could pass the Turing Test by imitating human conversation. Researchers at IHMC have developed and argued for an alternative view based on human-centered computing. In this view, the primary goal is to amplify, rather than replace, human abilities. It is one in which the AI system is seen as a cognitive prosthesis rather than a mechanical rival or competitor. In this new vision of our subject, the test of progress is the performance of the entire system comprising human and machines working together. This vision gives an entirely different technological emphasis, focuses on different medium-range goals, suggests different criteria for engineering success and provides a more rational basis for actual deployed applied AI technology than the traditional view. It places issues of human-machine interaction at the center of the subject, and focusses on achieving much tighter ‘fit’ between people and their artifacts.

Foundations of computation and information

The fields of theoretical computer science, mathematical computability theory and information theory are now mature, but still do not provide a fully satisfactory account of the basic natures of computation and of meaning, which would be adequate as a foundation for everything that falls under the category of ‘information technology’. Several active debates in contemporary philosophy of mind are tied to restricted and outmoded views of the nature of computation and information. Recent application areas such as self-aware computers, software agents, wearable robotic devices and the Semantic Web (all under investigation at IHMC) place new demands on our conceptual frameworks, but at the same time supply new intuition pumps. Researchers at IHMC continue to seek new theoretical models of computation and information that better reflect the engineering and social reality revealed by cutting-edge applications.

Probabilities, truth and meaning

Can we say that any computer system contains real meanings? There are philosophical arguments opposing this claim, but several IHMC researchers have argued for it. One of them says that real meaning arises from choosing rationally among several alternative meanings, using a process (called coherence or sense-making) that is essentially computational and can be implemented. In fact, IHMC researchers are implementing several varieties of these sense-making coherence-checking systems, and using them in realistic real-world tasks. The philosophical consequences of these ideas also are being pursued at IHMC, including finding clear criteria for when a machine can be said to believe or know something. Continuing interdisciplinary discussions between IHMC specialists in philosophy, logic, reasoning under uncertainty, linguistic, temporal reasoning, psychology, reasoning about preferences, and computational foundations all contribute to the intellectual climate and produce new insights.