Researchers at the Massachusetts Institute of Technology (MIT) devised an artificial intelligence system whose algorithms were trained by displaying horrifying texts and images of death found on Reddit, affecting the system and acting like a psychopath.
The researchers launched the first artificial intelligence system in the world with the name Norman Norman, in relation to the basic character in Alfred Hitchcock's Psycho film in 1960.
As part of the experiment, the researchers trained the Norman system using texts from Reed sites devoted to documenting and monitoring the horrific facts of death. Due to technical and ethical considerations, the Massachusetts Institute of Technology research team used descriptions and illustrations of images and did not use the real images of people dying to see if they would change these data and information from Behavior of artificial intelligence system.
While the basic rule of the Redet site groups - documentation of death - is that there must be a video of a person already dying in the publication being shared, and the transmission addresses must be descriptive and accurate enough to understand exactly what the subject content is like " Stabbed a young man. "
The researchers then subjected the Norman system to the Rorschach test series, a psychological test of personality and intelligence developed by Swiss psychiatrist Herman Rorschach in 1921. It tests what people see when they look at a series of ink spots representing typical abstract patterns and ask them to analyze or interpret them. Provided by participants psychologically for measuring possible thinking disorders.
A standard artificial intelligence system was subjected to the same Rorschach test as the Norman system to compare the results, as follows:
In one of the ink stains tests presented to the two systems, the standard artificial intelligence system saw it as "a close shot of a flower-decorated vase," while Norman saw it as "a man shot dead." In another example, the standard artificial intelligence system described the ink spot as "a black and white image of a small bird," while Norman described it as "a man drawn to the dough machine."
To answer the question, why did Massachusetts Institute of Technology create a psychologically artificial artificial intelligence system? The researchers said they are trying to learn everything about algorithms and that when things go so hard to blame the machine, as the data used to teach computer learning algorithms can greatly affect their behavior.
So when people talk about AI algorithms being biased and unfair, the culprit is often not the same algorithm but the biased data that has been fed.
The research team, which includes Bennar Yanardag, Manuel Ciprian and Iyad Rahwan of the Massachusetts Institute of Technology, said the study proved that the data used in the training of automated learning algorithms could significantly affect their behavior.
The study showed that the results of the experiment show that when the algorithms are accused of being "biased or unfair", the culprit is often not the same algorithm but the biased data entered into it.
&all knowledge with God .