I used the new Google NotebookLM to produce a podcast episode of my paper "SMARTA: Automated testing apparatus for visual discrimination tasks" (Vagell et al., 2018). The end product is very convincing. You really can't really tell that the 2 people in this podcast episode was AI generated. This podcast summarizes my SMARTA paper to about 16 mins and is a fun way to "listen" to the paper without actually having to read all that wall of text! I would say it is 99% correct with the exception that "colorblind" or "red-green colorblind" individuals don't see the world in black and white (grayscale) but instead cannot tell the difference between red and green (or any colors that is additive of these two colors). Instead, red and green looks brownish yellow to them (or to us, if you think about it!)
Check out the podcast episode below
Comments