Machine Execution of Human Intentions

Presentation PDF * Presentation Powerpoint presented May 20, 2014 at the AFCEA – GMU C4I Center Symposium: Critical Issues in C4I (George Mason University)

As military machines become ever more intelligent and more autonomous, the second order properties of man-machine coordination are going to become increasingly important. There have already been numerous expensive accidents due to the lack of communications of “intentions” from machine to man – and the ability of machines to process human intentions is getting closer to becoming reality. Further, machines need to become much better at recognizing and communicating anomalies if they are to avoid becoming vulnerable to both tragic accidents and intentional misdirection and “spoofing.” As battle is conducted at an ever-increasing pace in an increasingly complex environment, the interface and coordination of control between machine and man needs to evolve to take advantage of the strengths and support the weaknesses of each in order to maintain the edge and the safety that we need to be able to expect.

The Role of Metacognition in Creating Safe, Self-Improving Entities

Presentation Powerpoint presented May 9, 2014 at the University of Maryland Metacognition Seminar (College Park, MD)

Every AI system to date clearly demonstrates what Daniel Dennett calls “competence without comprehension”. While they might perform at super-human levels on specific well-bounded tasks, they are hopelessly brittle in the face of unexpected changes or anomalies. Previously, I, along with many others, have blamed “derived intentionality” for this state of affairs. In this talk, however, I will argue instead that it is “merely” a total lack of metacognition that is the sole (and traversable) hurdle which must be overcome in order to create artificial general intelligence (AGI). With the correct senses, tools and motivations, it should be possible to build a rudimentary metacognitive system that should not only dramatically ease the creation, maintenance and upgrading of current intelligent systems but pave the way for future intelligent software entities as well.

AAAI SSS14-03: Implementing Selves with Safe Motivational Systems & Self-Improvement

Organized by the Digital Wisdom Institute as part of the AAAI Spring 2014 Symposium Series March 24-26, 2014 (Stanford University, Palo Alto, CA, USA)

Artificial Intelligence (AI) and Artificial General Intelligence (AGI) most often focus on tools for collecting knowledge and solving problems or achieving goals rather than self-reflecting entities. Instead, this implementation-oriented symposium will focus on guided self-creation and improvement – particularly as a method of achieving human-level intelligence in machines through iterative improvement (“seed AI”).

Continue reading