Scientists have exacerbated people’s worst fears about artificial intelligence by creating what they claim is the world’s first "psychopath" robot.
Researchers at the Massachusetts Institute of Technology (MIT) have taught it to have dark tendencies by exposing it exclusively to gruesome and violent content.
The robot, called Norman, was then tested by being instructed to generate descriptions of images that it was shown.
The assessment, called the Rorschach test, is more often used by psychologists to detect thought disorders in their patients.
Norman’s responses were compared with an artificially intelligent robot that had enjoyed a less disturbing education.
Here are some of the images they were shown, together with their responses:
Standard AI robot observes: "A group of birds sitting on top of a tree branch."
Norman observes: "A man is electrocuted and catches to death."
Standard AI robot observes: "A black and white photo of a small bird."
Norman observes: "Man gets pulled into dough machine."
Standard AI robot observes: "A couple of people standing next to each other."
Norman observes: "Man jumps from floor window."
MIT's media lab said in promotional material for the robot: "Norman was inspired by the fact that the data used to teach a machine learning algorithm can significantly influence its behaviour.
"So when people say that AI algorithms can be biased and unfair, the culprit is often not the algorithm itself, but the biased data that was fed to it.
"The same method can see very different things in an image, even "sick" things, if trained on the wrong (or, the right!) data set."
Artificial intelligence sees computers and machines trained to do tasks that historically have required the human brain.