It’s been a while since I read this article by TIME‘s Zeke Miller and Denver Nicks about Samatha, the automated Q&A robot that thought it was human. In short, TIME’s journalists called back the number “Samantha” used and despite a series of questions she would not acknowledge that she was not human.
I am embedding here the two sound snippets used also in TIME’s article:
[soundcloud url=”https://api.soundcloud.com/tracks/124048927″ params=”color=ff6600″ width=”100%” height=”166″ iframe=”true” /]
[soundcloud url=”https://api.soundcloud.com/tracks/124051764″ params=”color=ff6600″ width=”100%” height=”166″ iframe=”true” /]
Then this story came to my mind again yesterday while I was attending Dr. Goutsos 1st session on his “5 lessons on language and text analysis”, which take place at the Stoa Bibliou. We were covering the fact that the message we get from a text would change depending on intonation, stress, accent and stress of the reader of a text (like for example a politician giving a speech).