Various types of knowledge can be retrieved from the Internet, but what about continuity of behaviors?
One more try?
I had an interesting conversation with Aaron Sloman and Benjamin Kuipers during AISB/IACAP congress. They doubt one can retrieve common sense about continuous happenings from the WWW resources. I have tried to extract behavioral steps (a'la Shankian scripts) before and I know it is tremendously hard but I wouldn't say it's impossible. I will definitely give it a try again and prove them wrong ;)
During the machine ethics session I was happy to talk to Mr. and Mrs. Anderson or Wendell Wallach (people who helped me much more than they realize) but it was Paul Oppenheimer who sent me a strong "no" signal to what I try to prove - that majority of people stay moral even online. I know they do not behave ethically all the time, but I don't think that trolls are in such numbers that could affect my algorithms. In another entry I will show that even without any semantic analysis a "majority choice" can agree with "offline" human subjects in almost 80%. Also Paul Bello has noticed that we are always weak to temptations - overeat, drink and smoke too much but this is, in my opinion, bad only in cases when other people are influenced and majority doesdescribesuch behaviors as bad / wrong.