The Computer That’s Infected With Schizophrenia

“All schizophrenia patients are mad, and none are sane. Their behaviour is incomprehensible.” —Kingsley Amis, Stanley and the Women

In A Nutshell

In an attempt to link schizophrenia with the excessive production of dopamine, researchers built a computerized neural network that would mimic the reaction of a brain overloaded with the substance. The network, called DISCERN, became quickly unstable, communicating with delusional stories and disjointed, nonsensical statements. It even answered one question by claiming credit for a terrorist bombing.

The Whole Bushel

Although true “artificial intelligence” might be forever out of reach, there is no denying that the technology is pretty close. When IBM’s Watson took on Jeopardy! champions Ken Jennings and Brad Rutter, it beat the humans soundly. Science fiction has been quick to flag the potential dangers of an artificially intelligent system, perhaps best represented by HAL 9000 in 2001: A Space Odyssey. A joint project by researchers at the University of Texas at Austin and Yale University has revealed just how dangerous computers could potentially be.

Wishing to better understand schizophrenia, the group built a computerized “neural network” which would mimic an overabundance of the neuroendocrine transmitter dopamine in the human brain. Their hypothesis was that excessive dopamine sends the brain into a kind of hyperdrive of overthinking that prevents people from having succinct, cogent thoughts that allow them to function. The schizophrenic mind, flooded with dopamine, creates false associations and assaults the brain with so much information at once that it is impossible to filter out what is important.

The researchers named their network DISCERN and began their experiments by programming the system to understand language the way a human does—not as bits of data to be stored, but as information that needs to be connected and cross-referenced. They fed DISCERN a series of stories until it essentially learned to “understand” them. Then, to simulate a human brain besieged by dopamine, they jacked up the computer’s learning rate.

DISCERN responded by acting schizophrenic, creating its own distorted set of connections. It responded to inquiries with jumbled, nonsensical strings of words and put itself in the midst of bizarre, deranged situations drawn from the stories it had digested, much like a person suffering from schizoid delusions distorts elements from real life into their own twisted narrative. In one instance, the computer claimed responsibility for a terrorist bombing.

These are certainly intriguing results, but they are far from absolutely proving the link between high dopamine production and schizophrenia. There are other substances, such as glutamic amino acid, which have also been tied to the mental illness. Glutamate is produced in excess when people take recreational drugs such as PCP (which has been linked to innumerable cases of temporary psychosis). If we can take anything away from the DISCERN experiments, it is that technology must be carefully monitored lest control slip out of our hands.

Show Me The Proof

US National Library of Medicine: Using computational patients to evaluate illness mechanisms in schizophrenia
University of Texas at Austin: Scientists Afflict Computers with Schizophrenia to Better Understand the Human Brain
NY Times: Daring to Think Differently About Schizophrenia

  • Hillyard

    And so it begins…

  • edzyl blane

    knowledgenuts – the greatest source of hollywood movie ideas

    • Nathaniel A.

      Why did you upvote your own comment?

      • lol

        Does it really matter?

        • Nathaniel A.

          No it does not, just curious.

          • lbatfish

            I tried upvoting one of my own comments once, after first doing it by accident. I was surprised that Disqus would let a commenter do that, because it wouldn’t be a difficult programming feat to prevent it. I did notice, however, that after I refreshed the page, my self-vote disappeared.

            Your comment was the first indication that I had, though, that anyone else could actually SEE a self-vote. And I also noticed just now that I, too, am seeing Edzyl’s, eight days after the original comment.

          • GerbilActs17

            the machines live…

      • edzyl blane

        I upvote every comment I see. There’s no limit, is there?

        • Nathaniel A.

          When it comes to your own, yes there is. Upvoting yourself is the Disqus equivalent of awkwardly high-fiving yourself after a joke you made that no one else gets.

          • CMatthew

            Wait your not supposed to do that?

          • edzyl blane

            Most of my comments are not jokes, but only questions to gather more opinions about a certain topic.

          • Nathaniel A.

            You misunderstood. Disqus= Comment, real-life= joke.

          • edzyl blane

            Ah. Ok. Just like liking your own status on facebook ?

          • Nathaniel A.

            Exactly.

          • edzyl blane

            Gosh, I hate those dweebs. Tell me, have I really turned into them?

          • Nathaniel A.

            So it would seem.

          • lbatfish

            IMHO, there’s still hope for you . . . IF you repent now!

  • Roger dela Cruz

    Plot twist: the computer was really the mastermind of a terrorist bombing.

  • daathsurfer

    That’s great. Damn; I don’t even know where to start. I think it’s fair to say what they mean by A.I. is actually synthetic cognition. In one experiment a slice of lab-rat neural networks were integrated into a computer in order to maneuver a small “car”. I think synthetic cognition is much closer than most people assume, especially considering what might be happening in classified projects. Perhaps the fabled “silicon life-form” will be invented by humans. There’s also some evidence to suggest we may be already living in a computer program. Physicist James Gates has discover error correcting code in super-string theory equations.Though intuition suggests someone must invent the computer first in order for us to be a computer program, that may not be the case. Considering our neural oscillations peak at about 100 hertz, we perceive our present about 80 milliseconds behind the true present. That compounded with the fact of wave-particle duality would deny substance to matter outside our perception; everything we perceive is resonating at our frequency. The fact matter can resonate at frequencies outside our perception demands an observer of the same frequency. Matter has no substance apart from our perception. The past has no substance outside our memory. The future is also non-existent in terms of substance. Thus, it is my theory the origin of the universe is in the future; about 80 milliseconds at all times. Between us and the present moment exists a cognitive computer operating exponentially faster approaching the limit of the electromagnetic spectrum. As it exists at higher frequencies than ourselves, we don’t have to invent it first in our subjective timeline; it exists closer to “now” than we do. We must invent it at some point in our subjective time, but it wouldn’t matter if it happened tomorrow or in a billion years. I came to this conclusion recently after watching a video about Pierre Teilhard de Chardin and his Omega Point cosmology. After that I learned Frank J. Tipler is a physicist with a theory similar (identical? I haven’t read his work) to my own.

  • Juilee Parkar

    I was reading the news and I saw this really interesting information

    You have perfectly created the intricate text in a very proven and straightforward fashion. I extensively love it.

    ANADROLE (ANADROL)

    Anvarol (Anavar) P-Var

    Crazy Bulk sale

    Dominant body building supplements

    Perfect body with strength only

    Crazy bulk Anavarol

    Crazy bulk discounts