Q & A about Cog

This is an interview of Joanna Bryson of the MIT AI and Cog laboratories, by Nicholas P. Drage of Nene College, Northampton, United Kingdom (92372349npd@nene.ac.uk).

  • Do you think that COG will ever be creative, in that he will conceive of ideas which were not expected of him?
    I hesitate to use the terms "creative" and "ideas". Cog will certainly do things we don't expect, it already does! In general, distributed intelligence has been very succesful at generating complex and interesting behaviors; Cog will hopefully do this even better.
  • Do you think that COG will ever be intuitive, in that he will seem to react based on something deeper and faster than what he is able to 'describe' - as in communicate to another sentinent, er, thing?
    I don't expect Cog to be able to describe anything, at least not any time soon. Most human intelligent behavior happens without being verbally described or planned in advance. It is probably an artifact of our consciousness that for some reason we consider such activities exceptional in particular situations.
  • I noticed that I called COG "he" in the last two questions without thinking, how do you think of COG - he, she, or it?
    I prefer "it", but that is probably mostly because I don't like gender biases. Cog certainly won't be able to reproduce in either gender-typical fashion.
  • When you say that "I like androids, they're my friends", how truthful are you being, and how ironic?
    I never said that, it was a Wired thing. Actually, I never said I used to think they were creepy, I may have said they were my friends meaning that I'm working on one... pencils are my friends too, when they work! (broken pencils are my enemies if I don't have a sharpener, you get the drift.) I personally thought the Wired article was kind of shoddy, though there were some reasonable points buried in it. The best article I've seen written about Cog is in the 20 May 94 issue of "Science".

  • page date: February 5, 1995
    page author: Joanna Bryson