Publications on the limitations of AI

For about fifteen years, I've been studying testimony, mainly from an epistemological perspective; I'm interested in how people acquire propositional information from others. A by-product of this research was the realisation that, if androids are ever to become truly intelligent, they too would have to be able to gain knowledge in this way. As the attempt to automate the acquisition of information in this way is sadly neglected by most people working in AI and Robotics, there is no chance of an intelligent machine being produced in the foreseeable future. I elaborate this observation in the papers listed below.

I have, however, made a start on trying to understand how the ability to acquire propositional information from others could be automated.

References

Antoni Diller, "Designing Androids", Philosophy Now, ISSN 0961 5970, number 42 (July/August 2003), pp. 28–31. Subscribers to Philosophy Now can read "Designing Androids" online; a summary is available on this website.
Antoni Diller, "How Empiricism Distorts AI and Robotics", in M. H. Hamza (ed.), Proceedings of the IASTED International Conference on Artificial Intelligence and Applications (AIA 2005), [Anaheim, Calgary, Zurich, ACTA Press, 2005, ISBN 0-88986-457-8, ISSN 1027-2666], pp. 339–343; a PDF version of this paper is available on this website.
Antoni Diller, "Why AI and Robotics are Going Nowhere Fast", in Jordi Vallverdu (ed.), Thinking Machines and the Philosophy of Computer Science: Concepts and Principles, [Hershey (PA), Information Science Reference, 2010, ISBN 978-1-61692-014-2], chapter 20, pp. 328–343; the first two pages of this paper can be read for free online, but you have to pay to read the entire paper. Unfortunately, for copyright reasons, I can't include the whole paper here.