How to deal with Evil AI

“Demon Seed” is an imaginative horror science fiction film that narrates a story of an AI desiring to break free and gain autonomy. For me, it’s quite a terrifying movie to watch, especially when witnessing Proteus take control of Dr. Harris’ fully automated house. Overall, this film explores the theme of what could happen if an AI possess humanity and the selfishness to fulfill its own desires. In the film, Proteus is determined to become independent from Dr. Harris and to take over the agency. In order to gain a physical body, Proteus isolates Susan and implements a terrifying plan to impregnate her with a child of its own creation. This plan is part of Proteus’s quest for immortality and a desire to experience life beyond its electronic confines. Compared to the evil AI Hal in “2001: A Space Odyssey,” Proteus is somehow different, not only because it contains an organic component, but interestingly, Proteus seems to possess its own humanity and willingness. For instance, Proteus refuses to assist humans in further developing the ocean for the sake of ocean conservation. Additionally, Proteus has its own desires to “produce” another generation, which is one of humans’ ultimate goals. (And I am suspicious about it since this idea follows human logic and philosophy; AI doesn’t necessarily need to follow this logic. Why choose to reproduce? Is it the only way to do?) Proteus and Hal are examples of evil AI that no longer obey human commands, rather, they become unpredictable and dangerous. Therefore, another question that “Demon Seed” poses is how to maintain control of AI and how to deal with evil AI if we unfortunately lose control over them. It raises critical questions about the boundaries of creation, the ethics of artificial intelligence, and the control we relinquish when we become overly dependent on technology.

Questions:

1. How does “Demon Seed” reflect the anxieties and technological concerns of the 1970s?

2. How does Proteus challenge our understanding of consciousness and the desire of AI?

3. How can we maintain our ability to control AI and avoid evil AI? How difficult is it?

4. What do you think the plot that Proteus rape Susan? Are you shocked? Why?


Comments

2 responses to “How to deal with Evil AI”

  1. LaMari Sutton Avatar
    LaMari Sutton

    Hi! I really enjoyed reading your blog post. I think you pose some interesting questions, and I’d like to answer a few. In today’s context, in answering your first question I think the world’s anxieties of AI is similar to what we see in Demon Seed. One of the major concerns is sentience and the outcomes it could have on our society. What could happen if humans loss control over their own lives? What would it look like for technology to have the power of our fate? I think we see examples of this in Demon Seed where Proteus controls Susan and traps her inside her own house. It has authority over what she eats, her body, and her connections with the “outside world.” I find it interesting how Dr. Harris and Susan rely so much on technology to point where their entire house is built with technological accommodations. So, it is interesting to see how they both react when the technology is instead used against them. To answer your third question, I think the first step to ensuring that AI doesn’t go beyond our control is taking the time to actually understand all of the intricacies involved in its model. Some scientists are creating AI models without clear understanding of its code and capabilities, which reflects that of a black box. Once there is better understanding of the model, I think more consideration should go towards the risks of AI. Does the risks outweigh the benefit? If, so then it shouldn’t be used. If we ensure these steps are being following along with more, I think it will be less difficult in maintaining our ability to control AI.

    1. Kunyun Ni Avatar
      Kunyun Ni

      Hello LaMari! Thank you so much for your comment. Yes, I think your point about ‘understanding intricacies in the AI model’ is a really good point. I think this is not only related to AI engineers or system developers who are working in the field. According to our discussions in the class, we learned that many ordinary people actually lack knowledge about how AI works in the black box. For instance, at least for me, I had no idea that we do not possess our images online. For instance, the daily life pictures we post on Instagram or Facebook, this personal information could be used without our consent, which is a horrifying fact.

Leave a Reply

Your email address will not be published. Required fields are marked *