Norman, the Psychopathic AI.
AI-Powered Psychopath
Scientists at the Massachusetts Institute of Technology unveiled the first artificial intelligence algorithm trained psychopath Norman, like its namesake Hitchcock’s Norman Bates, it does not have an optimistic view of the world.
Norman performs image captioning, “a deep learning method” that allows AI to generate text descriptions for images. However, the team exclusively exposed Norman to violent and disturbing images posted on a subreddit purely dedicated to death; affirming their predictions regarding the affect on an AI when trained solely on data from “the dark corners of the net.” When a “normal” algorithm generated by artificial intelligence is asked what it sees in an abstract shape shown through inkblot drawings, it chooses something cheery: a group of birds sitting on top of a tree branch. Norman, on the other hand, sees a man being electrocuted.
And where “normal” AI sees a couple of people standing next to each other, Norman sees a man jumping from a window.
These abstract images are traditionally used by psychologists to help assess the state of a patient’s mind, in particular, whether they perceive the world in a negative or positive light. Norman’s view was unremittingly bleak — it saw dead bodies, blood, and destruction in every image.
Close by Norman, another AI was prepared on more ordinary pictures of felines, winged creatures, and individuals, which saw unquestionably more bright pictures in a similar theoretical smears.
“The way that Norman’s reactions were so a lot hazier delineates a cruel reality in the new universe of AI,” said Prof Iyad Rahwan, a memeber of the three-man group from MIT’s Media Lab which created Norman, “Data matters more than the algorithm, It features that the data we use to train AI is reflected in the manner the AI sees the world and how it perceives.”
Artificial intelligence is all around us these days and AI is already being deployed across a wide variety of industries, from personal digital assistants, email filtering, search, fraud prevention, voice and facial recognition, and content classification, to name a few. But Norman is just a gist of AI trained on bad data that itself turned bad.
It is not mathematically possible to create fairness, moreover, bias is not bad for machine learning, all it implies is the regularities picked up by the machine while training.
What is worrisome is some programmer’s malice behaviour to deliberately choose to hard-bake the machine with inappropriate data. Punctuating this process of creating AI requires clear oversight and great transparency.
AI psychology
Psychology is the study of mental processes and behavior of individuals. Artificial Psychology is then the study of the mental processes of an Artificial Intelligence System (AIS) similar to humans. It is about the artificial cognitive processes required for an artificially intelligent entity to be intelligent, learning, autonomous, and self-developing. There is also developmental psychology that considers how an individual adapts and changes during different developmental stages and what is appropriate to consider of a human based on development. The AIS must mimic human processes in order to be intelligent.
This new era of “AI psychology” would take the form of regular audits of the systems being developed, rather like those that exist in the banking world already. There is a growing belief that machine behaviour can be something you can study in the same way as you study human behaviour.
Norman is a great way to start an important conversation with the public and businesses who are coming to rely on AI more and more. We are teaching algorithms in the same way as we teach human beings so there is a risk that we are not teaching everything right.
