I was wondering if anyone can think of any examples of integration of audio-visual inputs, or other modalities, failing or resulting in undesired outcomes in artificial cognitive systems?
I'm interested in any issues, or artifacts, that have arisen when different information about the same event have been integrated in such a system.
Published, or discussed in print online, or even a TED talk would be great sources if you know of any. Anecdotal evidence of such scenarios that you may have experienced are also of interest.
Thank you all for your time.