I set myself a challenging task when I put this topic on my list of blog posts for 2026. Examples of presumably sentient or conscious artificial intelligences in science fiction are plentiful and varied. There are harbingers of dread and hope aplenty.
Yes, the list is long and detailed (List of fictional computers – Wikipedia) and includes benevolent and malevolent examples. It seems, at the risk of being obvious, the various authors and screen writers constructed the AI and its behavior for what was needed to tell the story or address the authors’ themes.
One of my favorite examples is The Mechanical Brain from Edgar Rice Burroughs’s Martian Series: Swords of Mars, published in 1934. The brain was used to control a warship, and telepathically, which proved the villains undoing when the hero, John Carter, takes command and saves the girl.
(Image Credit: https://i.pinimg.com/736x/80/8a/0f/808a0fce4dbb78146fc0ca61750f0d85.jpg)
The list continues in books and films to the present where we meet Lilypad, the sentient tablet computer in Toy Story 5.
(Image Credit : https://www.laughingplace.com/disney-entertainment/toy-story-5-jessie-lilypad-clip/)
Given the variety in fictional AIs that has come before, let’s consider what might be the attributes of a sentient AI. To simplify a complex question, I settled on considering four, separate, but interrelated concepts that we can use to discuss the question: (Note: these are fairly rudimentary, accessible definitions from “googling” the word)
Sentience, which can be defined as the fundamental capacity to have subjective experiences, specifically the ability to feel pleasure, pain, or emotional tones.
Consciousness is the state of being awake and aware of one’s surroundings, internal thoughts, or having an “internal observer”.
Intelligence might be the most concrete/functional as defined as the computational or cognitive ability to acquire, process, learn, and apply information to solve complex tasks and adapt to new environments.
Self-awareness is the ability to reflect on one’s own mental states — to know that you are thinking, that you are feeling, and that you exist as a distinct entity.
When we think of AI or use any of the popularly available AI tools, it’s likely the intelligence attribute we encounter. We can get our questions answered, we can find information, we can have text composed, etc. But I’d bet we don’t think of Claude, for example, as being able to experience the world. Does the AI writing a review of a recipe or a restaurant feel hungry, is it horrified if it’s writing a term paper on the holocaust. In this latter instance, it might be able to summarize the horror written by human authors, but would it experience their reaction. That would imply then that these AI models are not sentient.
That brings us to consciousness. In his seminal work on the function of the hippocampus, Dr. Doug Burman (Note 1) suggests artificial intelligence and purpose-built computers, programmed to perform these functions, can regulate sensory input, thoughts, and responses to the environment, but they lack the ability to express emotional responses or create original thoughts. This would argue then that even the most sophisticated computer/AI is not conscious as humans would experience it.
Now to the question in the title of this post: Do they know they are artificial? To answer yes, this would require that the AI is self-aware, that it recognizes itself as a unique entity, separate from others and that it can “examine” its own thoughts. In our current understanding, the answer would be no, the AI lacks the emotional framework or consciousness required for self-awareness. By default, since an AI would not experience itself as a distinct entity, it would lack the awareness to know if it was artificial.
To wrap up though: could we or will we develop an AI that would emulate these four attributes and become basically an artificial human? I would suggest the better question is: Why would we? Why would we need to re-create what we already have in abundance?
To work better or faster? They don’t need to be conscious or self-aware for that. To perform tasks that are in short supply, such as care givers? It might make them easier to relate to, but they don’t need to “care about” their charges, do they?
In my next post, I’ll pick up the topic of work: Robots and Human Activities: What Will We Do with All the Free Time?
In the meantime, the question then remains: Why would we build AIs and by extension robots, that precisely emulate human behavior or abilities?
Give it some thought and let me know what you think.
Thanks for stopping by.
Note 1: Burman, Douglas D. (2023) Normal Alert Consciousness – A Central Executive Model of Hippocampal Function. http://www.intechopen.com/books/normal-alert-consciousness-a-central-executive-model-of-hippocampal-function, published by IntechOpen
Your distinction among intelligence, sentience, consciousness, and self-awareness helps separate two questions that are often run together. An AI could represent and report the proposition that it is artificial without there being anything it is like to be that system. The harder issue is what would make such a report knowledge rather than a programmed or learned self-description. The same challenge seems to apply to humans under naturalism: an evolutionary story may explain why a self-model produces consciousness-talk without yet explaining why that judgment reliably tracks actual subjectivity.
This video explores the meta problem of consciousness proposed by David Chalmers and argues that naturalistic processes couldn’t have given us knowledge of consciousness
This is my video: https://www.youtube.com/watch?v=WJVvZNi0Fi8
Would the central-executive model you cite predict some marker that makes a first-person consciousness report truth-tracking, or does it explain only how a system generates the self-description?
PP – interesting comment on the difference between the robot knowing it’s artificial and the creators programming that response/knowledge. Appreciate it.
And – my friend who’s a neuroscientist who helped me start to think about this question, had the following comment (used with permission):
This post with the YouTube video represents one end of the spectrum of ideas on consciousness that have been put forth… and although it appears rational on the face of it, I see several problems with it.
First, there’s something called “neural correlates of consciousness (NCC),” markers of brain activity associated with conscious sensations, feelings, etc. In the YouTube example, if someone closes their eyes and imagines a green square, sensory areas of the brain will in fact become active as if he was actually looking at a green square. Furthermore, if a green square was competing with other sensory stimuli for his attention, there are brain areas that would respond to the green square only when he was consciously aware of it — hence the term, “neural correlates of consciousness”. Unlike the claim in the posted video, the imagined green square does not exist outside of the person’s brain, but you can identify its existence based on brain activity.
Conscious awareness of sensory stimuli (for example) does not merely result from the brain’s electrical activity… it results from specific parts of the brain that are active and transmit information to other functionally relevant areas. Conscious awareness is associated with a ‘self’ — and there is a part of the brain that specifically identifies our physical self.
My central executive theory of consciousness says that in the “normal alert” state of consciousness (one of many), the hippocampus is at the center: sensory signals must reach here for conscious awareness, and signals from the hippocampus are sent to other areas of the brain to act upon these signals for context, emotional feelings, to initiate movements, to access relevant memories for making decisions, even for creative endeavors. There is direct evidence from fMRI and other neuroimaging for some of these claims; there is also indirect evidence from problems identified with hippocampal connections in neurological disorders that impair various conscious functions.
The YouTube video tries to logically prove the idea that a naturalistic explanation can’t explain consciousness, thus leaving only a supernatural explanation. The argument has logical fallacies — but even if we are only left with a supernatural explanation, how can the supernatural (“invisible dragons”) act on my brain to produce intentional movements? We know the brain is necessary for intentional movements, and executing other decisions we consciously make.
The YouTube video claims that the brain cannot create a conscious self. If that’s true, the supernatural, conscious self must have some means to influence the brain — you can’t have it both ways. (Unless you agree with Daniel Dennett, who believed there is no such thing as consciousness, only the illusion of consciousness.) This YouTube video does not address this issue.
The best reason for building robots to emulate humans might be to provide companionship and sustain cultural traditions over extended periods of time, such as during intergalactic travels that require several generations. With replaceable parts, the robot might provide a continuing source of comfort, knowledge, and culture that otherwise might be lost within a few generations.
The risks, however, would be high. A parent can provide guidance but can’t guarantee what kind of adult their child will grow into; the same risk would hold for a human-like robot. The robot would have to learn preferences, as humans do, and find a sense of self… and its evolution of likes / dislikes, loyalties, and interests could not be predicted. Perhaps it would become an outlaw. Outcast humans can be convinced to conform through societal norms and punishments; how can a human-ish robot who will outlive us all be convinced to conform?