Scientists have created the world’s first artificial intelligence (AI) to display psychopathic tendencies – and named it after Norman Bates.
Exposed to some of the dark corners of the internet, the algorithm continues to come up with deeply disturbing interpretations of the images it is shown.
Researchers at Massachusetts Institute of Technology (MIT) created the algorithm nicknamed in tribute to the character in the 1960 Alfred Hitchcock horror film ‘Psycho’.
It was trained using gruesome picture captions from Reddit, causing it to associate objects with death.
They produced what has been labelled the “world’s first psychopath artificial intelligence”, which sees horrific images in everything.
MIT scientists Pinar Yanardag, Manuel Cebrian and Iyad Rahwan are behind the project, They gave Norman a Rorschach inkblot test to explore what it sees.
Disturbing descriptions produced by the AI include “a man is electrocuted and catches to death”.
A standard artificial intelligence algorithm instead saw “a close up of a vase with flowers”.
While the Norman project immediately brings to mind psychopathic robots from the ‘Terminator’ movie franchise, the researchers say there is validity in showing that machines are not inherently biased and that the people inputting data can significantly alter their behaviour. There have previously been examples of algorithms showing disturbing tendencies when exposed to social media.
In 2016, Microsoft apologised for creating an online “chatbot” that quickly turned into a Holocaust-denying racist.