As we charge into the future, technology is advancing faster than ever before. In Susan Schneider’s Artificial You, she predicts that artificial intelligence (AI) will be able to “‘carry out most human professions at least as well as a typical human’ within a 50 percent probability by 2050, and within a 90 percent probability by 2070.” With these accelerated advancements, conversations surrounding the potential future of AI are more necessary than ever before.
When speculating about the future, researchers have theorized about the possibility of uploading a person’s consciousness to a computer, but how could this be possible? William Cheshire addresses this idea in his article “The Sum of All Thoughts.” Cheshire first likens the human mind to a computer saying, “the brain thinks with ‘wetware.’ After all, quipped a fictional neurosurgeon on the television drama Three Pounds, the brain is just ‘wires in a box.’” This parallels Schneider’s views surrounding the characteristics of the human mind when she states, “the brain is an information processing engine and [all] mental functions are computations.” As Cheshire continues, he solidifies this idea, referencing the idea of substrate independence, or the idea that mental states can occur on a broad class of physical substances. For example, think about the way your brain is able to store and process information. These processes could theoretically be replicated by a silicon chip mimicking the function of a human brain.
Schneider combats this substrate independence theory when she addresses biological naturalism. Biological naturalism “claims that even the most sophisticated forms of AI will be devoid of inner experience.” In Schneider’s view, Cheshire would be considered a “techno-optimist,” referring to the belief that consciousness is computational, and therefore complex computer systems will have experience. Cheshire reveals his techno-optimism when he states, “Personal identity, memories, likes and dislikes, loves and fears, beliefs and aspirations—consciousness itself—would be reframed in a substrate of silicon, copper, plastic, and glass, enclosed perhaps in a polished aluminum pseudo cranium.” This leads Cheshire to his argument about something he calls “uploading.” In traditional computer engineering, this term refers to the transition of data or information from one computer system to another higher-level computer. It should be noted, however, as Cheshire explains, the process of “uploading” a human brain is much more complex, and we currently do not have a deep enough understanding of the human mind to do so. Cheshire writes, “There is still a great deal about how the brain works that neuroscience has not deciphered. Although neuroscience has shed considerable light on the functions of the brain, it lacks the ability to explain the phenomena of consciousness, personal agency, conscience, moral responsibility, the continuity of identity over time, or human purpose.” In other words, we cannot upload something we do not fully comprehend. Though Cheshire’s subscription to the theory of substrate independence may suggest the possibility of mind uploading in the future, we are still unable to put a thumb on what makes us feel things or the reasons we have morals. So, where does that put our current understanding of consciousness? And is consciousness something AI could attain?
Drew McDermott engages in these arguments regarding AI consciousness in his paper “Artificial Intelligence and Consciousness.” McDermott begins by introducing a concept he calls the “Moore/Turing inevitability” argument. He coined the term because of the concept’s reliance on Moore’s Law. Moore’s Law predicts exponential progress in the power and intelligence of computers. McDermott’s concept first addresses our inevitable encounters with AI. McDermott uses the example of a skilled poker player machine: “If we had a talking robot that could play poker well, we would treat it the same way we treated any real human seated at the same table.” That is to say, if a machine can adequately imitate the actions of a human, humans are more likely to address the machine as if it were human. In Schneiders’ writing, she discusses a similar idea. She says, “A sophisticated AI could solve problems that even the brightest humans are unable to solve, but would its information processing have a felt quality to it?” In other words, despite an AI’s capabilities, the information being processed to enable these abilities would be anchored in mathematics and far from the human experience of emotional, feeling-based decisions. On these grounds, McDermott and Schneider agree that although AI can act as though it is conscious and make decisions that imply consciousness, the process involved is an algorithm.
Where McDermott and Schneider disagree is in the discussion of what the capability of mimicking consciousness makes AI. A quote from McDermott’s essay sums up his stance on AI consciousness quite well: “To be conscious is to model one’s mental life in terms of things like sensations and free decisions. It would be hard to have an intelligent robot that wasn’t conscious in this sense, because everywhere the robot went it would have to deal with its own presence and its own decision making, and so it would have to have models of its behavior and its thought processes.” McDermott believes that to be conscious is to have the ability to use free decisions when modeling a mental life. This would be essential for an intelligent robot to exist and operate in the human world. That is to say, if, in the future, AI lives among us by McDermott’s standards, these machines would be considered conscious beings.
Schneider, on the other hand, views consciousness to be less literal. It goes beyond just free decisions, which is why she introduces the concept of “synthetic consciousness.” Schneider theorizes that machines can become conscious because we know the brain is conscious, and we could build an exact replica or isomorph of the brain. This method implies that in order to be conscious, machines must adopt every single function of the brain. It is as complex as it sounds. As Schneider further explains, given our current knowledge surrounding neuroscience, it seems impossible to create the model previously suggested. Of this issue, she states, “Again, I take a wait and see approach. But the possibility that advanced intelligences outmode consciousness is suggestive. Neither biological naturalism nor techno-optimism could accommodate such an outcome.” Schneider believes we can only look to the future for the answer to this question, while McDermott presents a more optimistic view, arguing that perhapss AI consciousness is closer than we think. Suppose humans will share life with AI in the future; we must understand what this means, the possibilities, and the potential threats this poses to the human race as we know it.
This notion of cohabitating with AI is brought to question in Elizabeth Kearney’s “Artificial intelligence in genetic services delivery: Utopia or apocalypse?” Kearney sets the stage for our current technological environment, noting movies such as Space Odyssey, which have painted AI as something to come in the future. However, Kearney argues that the age of AI has already arrived. Kearney credits this arrival to systems such as search engine algorithms, digital personal assistants, facial recognition, chatbots on retail websites, and more. In fact, Kearney believes that we have become so accustomed to AI and its function in our lives that we don’t even realize its prevalence. In contrast, Schneider does not acknowledge these forms of AI and instead views AI as the stuff of the distant future. She goes on to state that humans are just “an intermediate step to AI, a rung on the evolutionary ladder.” Though some optimists find this new step of evolution to be groundbreaking and exciting, Schneider notes the idea of the ‘control problem.’ She writes, “AI could be our greatest invention and our last one. This has been called the ‘control problem’—how we Earthlings can control an AI that is both inscrutable and vastly smarter than us.” In other words, how can we put ourselves above something that outperforms us in every aspect? Schneider goes on to explain the flaws of programming morals into AI; she claims that as AI becomes more intelligent, any such programming could be overruled. It is universally believed in the realm of computer technology that a machine could bypass and override safeguards, such as kill switches, which could potentially threaten human life. This seems to be the biggest potential threat of AI in our modern world: control, or lack thereof. Through the systems Kearney presented, it is clear that AI outperforms humans in a significant amount of real-world applications. But anything further than a non-sentient algorithm could result in danger or, as Schneider explains, the end of the human race as we know it.
Works Cited:
Cataleta, Maria Stefania. “Humane Artificial Intelligence: The Fragility of Human Rights Facing AI.” East West Center, January 2020.
Cheshire, William, Jr., MD. “The Sum of All Thoughts: Prospects of Uploading a Mind to a Computer,” Ethics & Medicine 31, no. 3 (Fall 2015): 131-141.
Gamez, David. “Machine Consciousness.” In Human and Machine Consciousness, 1st ed., 135–48. Open Book Publishers, 2018. http://www.jstor.org/stable/j.ctv8j3zv.14.
Kearney, Elizabeth, et al. “Artificial intelligence in genetic services delivery: Utopia or apocalypse?” Journal of Genetic Counseling, Vol. 29, No. 1, February 2022. Pages 8-17.
McDermott, Drew. “Artificial Intelligence and Consciousness,” Cambridge Handbook of Consciousness. Cambridge: Cambridge University Press, 2007. Pages 117-150.
Moore, Gordon E. “Cramming More Components onto Integrated Circuits,” Proceedings of the IEEE, Vol. 86, No. 1, January 1998. Pages 82-85.
Nahmias, Eddy, Corey Hill Allen, and Bradley Loveall. “When Do Robots Have Free Will?: Exploring the Relationships between (Attributions of) Consciousness and Free Will.” In Free Will, Causality, and Neuroscience, edited by Bernard Feltz, Marcus Missal, and Andrew Sims, 338:57–80. Brill, 2020. http://www.jstor.org/stable/10.1163/j.ctvrxk31x.8. Schnider, Susan. Artificial You. Princeton University Press, 2019.
