Artificial intelligences fail to identify optical illusions in images created by other AIs – so these images could form the basis of a new kind of CAPTCHA test
Comments
Log in with your Bluesky account to leave a comment
Fascinating! Ever since our GECCO 2013 study I suspected machines to eventually outperform humans in distinguishing between human and machine behaviour. This is why Turing learning could in principle produce machine behaviour considered human. https://en.m.wikipedia.org/wiki/Generative_adversarial_network#History
Comments
https://en.m.wikipedia.org/wiki/Generative_adversarial_network#History