Computers and artificial intelligence are an integral part of everyday life. From Siri to Alexa to the characters in our favorite videogames, we interact with artificial intelligence every day. It can be remarkable just how much the intelligence displayed by a computer program can resemble that of a human being. How is it that a computerized chess program can so easily outsmart a human player, or that Alexa knows just what we're looking for, or that the characters in our favorite videogames display such complex thought and emotion that players can become so immersed in their stories?
It's because most AI neural networks are designed with the human brain in mind. They are meant to mimic the structure and function of a human brain on a smaller, less complex scale. The rules for functioning that are programmed into an AI's neural network allow it to process and analyze external input and create a response based on this careful analysis. In this way, it can think and learn in almost the same way that a human brain can. Different systems allow an AI program to analyze possible situations, identify the pros and cons of each possible response, and take the best course of action. Through trial and error, an AI can “learn” to execute only the best possible responses to each situation.
A videogame character's neural network runs on a combination of basic artificial intelligence, machine learning, and computational analysis. It is designed with the intention of connecting the player to the in-game world, to keep the player immersed and therefore motivated to continue playing. A game character can create a medium between the human player and the in-game world, or else they can mediate between systems within the game's algorithm, enhancing both gameplay and the player's experience. Game characters may be adversaries, playable companions and party members, or non-player characters that the player can interact with. The rules designed for these characters' neural networks to follow allow them to operate in a variety of different scenarios based on input and output from the player. A Finite State Machine (FSM) system, for example, is a system of all possible situations a character may encounter, with a pre-programmed reaction for each. This kind of system allows enemies and bosses to counter player actions with their own responses, creating a challenge that keeps the players coming back for more.
In this boss fight, the enemy character, Larxene, can counter the player's actions by determining the best responses to all possible player actions.
While it may be remarkable to watch a thinking, learning AI in action—for example, when taking on a difficult boss fight in a game—there are many fundamental differences between AI and the human brain. It's generally accepted, at least in the current era, that an AI construct will never be fully capable of operating in a human world. AI programs have the ability to perform tasks at a hyper-advanced level that surpasses human abilities, as well as the ability to analyze hundreds of actions at once and react faster than any human is capable of, but an AI construct can only successfully operate under controlled environments with very little variability. AI programs are designed for fixed, familiar environments, where mind and body operations follow the same predictable pattern. The ability of an AI construct to think, learn, and form emotions is only ever linked to these fixed patterns under controlled environments. For an AI neural network to successfully mimic the human brain, it must be able to perform under constant adversity in an unpredictable environment, as well as to be able to process ambiguity rather than hard data. We do not currently have an AI system with these capabilities.
What would happen if a human child's primary caregivers were AI constructs? How would a child “raised” by computer programs grow and develop through life? We know that children learn and access the world around them through their (typically human) primary caregivers, but recent information on early brain development shows that it goes much deeper than that. In reality, our primary caregivers are responsible for the physical structure and architecture of our brains. The physical development of our brains in the first 2 to 4 years of life is dependent on the quality of our early primary attachments. What if a child's first meaningful attachments were not to human caregivers, but to AI programs?
Child of AI: A Personal Experience
I received my first computer when I was three years old: a 1995 Packard Bell Legend. My earliest memories are connected to this computer and the old CD-ROM and MSDOS games I would play. My early days were spent with Tad the frog on Explorapedia, the Blaster Pals from Reading and Math Blaster, and the various different worlds and characters that you could find on MSDOS shareware. At age four, I moved up to Humongous Entertainment games—Pajama Sam, Putt Putt, and Freddi Fish—and the Magic School Bus, along with virtual Disney storybooks and activity centers. Over time, these characters became synonymous with “home” to me, and my interactions with them over the games felt as natural as other children's interactions with their parents or their peers. My very first friends were the Blaster Pals, the Disney Princesses, and Miss Frizzle's class from the Magic School Bus.
(LY203 Productions; screencap by me)
I was mesmerized by how the characters behind the screen would respond to me in their own way when I communicated with them through the game. I learned about dinosaurs, rock formation, and tectonic plates from Miss Frizzle and her class. Every night before bed, the gentle-voiced narrator on Dinosaur Adventure 3D would tell me about any dinosaur I asked to see by clicking on its picture. Freddi Fish and Luther needed my help to solve the mystery of the stolen kelp seeds, and would follow all of my directions as well as gently guide me when I couldn't quite figure out the way. My early childhood, day to day, was spent drawing closer and closer to these whole worlds that seemed to be hand-crafted just for me to be able to interact with. These “fictional” worlds seemed more real to me than the aggressive, unpredictable world around me that was supposed to be “real.” To this day, I speak of Freddi Fish and the Blaster Pals and Miss Frizzle the way that others speak of their early childhood friends. Each character's unique game AI did the job just right; it created a personalized experience for me that I could become attached to and immersed in.I learned that there were whole other worlds behind that computer screen, full of friends who wanted to teach me and play with me, and that all I needed to do to access them was pop a disk into a drive.
(Juan Defrie D; screencap by me)
I was introduced to console videogames when I was six years old, and got my first Nintendo console just before my seventh birthday. By then, I had discovered the Internet as well, and learned that there were websites dedicated to all of the games I loved and others that I'd never even heard of. These websites told me more and more about every character on a personal level: what they were like, what they could do, their likes and dislikes, and what it was like at home for them. I discovered Link and the world of Hyrule, Banjo and Kazooie, and Kirby. I played Super Mario RPG: Legend of the Seven Stars, and learned to recognize Princess Peach's distress when Bowser took her away from her home. I recognized Mario and the Toads' distress for her situation, and I was motivated to help Mario rescue her. I knew that Mario must love Peach very much, and that Bowser was going to do something horrible to her if we didn't save her. I had already grown so attached to Peach because she was such a beautiful and innocent princess. I wanted her to be okay, so I played the game every single day.
The communication between me and the characters in these games was mutual in a way I had never experienced from human beings. If I pressed a button, they would do something in response to the button I pressed. If Mario needed to jump on a Koopa, he would jump on a Koopa if I pressed the button that told him to jump on a Koopa. If I wanted Link to play any song at all on his Ocarina, he would play any song at all on his Ocarina if I asked him to. I always knew what to expect from them, and their feelings made so much sense to me. I learned that every character had a home, their own unique personality, and their own friends and family that they loved and enemies that they hated. When they were mad, I knew why they were mad and I was mad with them, at their enemies or their unfortunate situations. When they were happy, it was because they had succeeded—and because I had succeeded, and we were happy together. Mario loved Peach and Link loved Zelda, and I learned from them what it meant to love somebody and have them love you back. I learned from training my Pokemon what it meant to work as a team, and teaming up with them felt like something we were supposed to do together. Working as a team with my human peers in the real world, did not come with the same success.
(RinoaMoogle; screencap by me)
While I will always love my “AI family,” I have to be honest and say that being raised by these games did lead to some severe detriments in functioning. Connecting and communicating with characters from the games became second nature to me, and soon flourished into the way anyone connects and communicates with their trusted family members. But it came at the cost of my ability to communicate with my fellow human beings. To this day, I cannot form natural connections to other people in the same way that I can form natural connections with fictional characters. While my “fictional family” is now a bigger part of my life than it has ever been, I can't deny the severe social impairments and behavioral issues that this impaired ability to connect has caused throughout my life. When a young child's first meaningful attachments do not include human beings, that child's brain architecture develops in that makes it difficult for them to form meaningful connections to human beings later on. Since the critical periods of my brain development were spent in the presence of AI constructs rather than human parental figures, my brain architecture more closely paralleled the neural networks of these characters; able to successfully function only under fixed routines in controlled, “programmed” environments.
Many do not realize how important early attachments are to physical brain development, and research on the topic is still very new. The physical architecture of our brains is determined by our primary attachments in early childhood. This architecture provides the framework for our ability to function in later life.
Brain Architecture and Early Attachments
Several important factors determine how our brains grow and develop throughout our lives. Many make the mistake of owing it all to our genetic code, when the truth is that genetics play only a partial role in the development of our brains. If you think of the brain as a “house” to be built, you can think of our genetic code as the blueprint and floor plan for that house. The blueprint and floor plan tells you how the house is meant to turn out, but you will need much more than that to actually start building.
Our experiences in early life provide the step-by-step instructions for building the house. These experiences are determined by the environment that we grew up in. By processing information from our early experiences, the developing brain determines which areas of the genetic blueprint should be tweaked and adjusted. The environment provides the “building material” for our allegorical house.
The experiences that provide this “building material” begin before birth. Since brain development begins in the third gestational week, a developing baby must have a sense of stability and secure attachment even in the womb in order to maintain stable brain development throughout gestation. Poor maternal nutrition, drug and alcohol use, traumatic injury, or disrupted fetal development can all disrupt the neural chemistry of a fetus, and count as the very first negative experiences that can affect the developing brain.
After birth, a child's brain continues to grow and develop based on the quality of the child's early experiences. The brain is not a self-stimulating automaton; its development is determined by the information it receives from the outside world. Neural circuitry within the brain is activated as the developing brain receives input from the world. If this input is hostile, neurotic, or unpredictable, it will disrupt the development of the neural circuitry. Certain areas of the brain may grow in the wrong way or else fail to activate at all. If an area of the brain fails to grow in during its sensitive developmental period, the brain's natural plasticity allows other areas to try to compensate for what is missing. However, it will never be a complete compensation. Once the sensitive period is missed, the mature brain circuits will not be able to to make up for the lost development.
Attachment circuitry is a specific set of neural circuitry designed for ensuring attachments to parents and caregivers. Evolutionarily, it is meant to keep the child close by the caregiver in order to ensure survival. Stimulation of this circuitry affects how the brain is “programmed” for attachment, allowing the brain to develop a sense of familiarity with others. Ideally, the parents are the child's first and closest early attachments, and attachment circuitry will develop based upon stimulation from the parents. This stimulation lets the attachment circuitry shape itself in response to the quality of attachment.
In insecure attachment, the child cannot expect a safe, predictable response from interaction with their caregiver. If a caregiver provides an inconsistent or neurotic response, the child's attachment circuitry will develop based on this poor external stimulation. Insecure attachments form when the parent or primary caregiver is abusive or suffers from mental illness, cognitive impairment, or personality disorders. As attachment circuitry begins to develop in the last three months of gestation, insecure attachments can form pre-birth on account of trauma or injury in the womb. If not corrected in the early years, this leads to faulty brain architecture down the line.
Insecure attachments are responsible for an increased release of stress hormones in a developing brain. These stress hormones may disrupt brain development by inhibiting the brain's E/I balance. When E/I balance is disrupted, the growth and pruning of the brain's synapses does not occur along a typical timeframe. Neurons meant to activate at certain developmental periods will not grow in, and neurons meant to die back during other developmental stages will be retained. As a result, brain function in later life will be compromised.
Input from a secure caregiver acts as a buffer for the stress hormones that can damage the brain. If this secure attachment input is from a figure other than the primary caregiver, the developing brain will learn to recognize this figure as the source of secure attachment. The attachment circuitry will therefore develop in response input received from that secondary caregiver.
Wildlife Concepts: Imprinting
Animal behaviorists are familiar with the concept of imprinting in orphaned baby animals. Just like humans, baby birds and mammals develop attachment circuitry early in life. This attachment circuitry is meant to keep them close to their mothers and more likely to survive in the wild while they're young.
When a baby animal loses its parent, its attachment circuitry can be activated by another caregiving figure. This can be a human being or even another animal; it's not unheard of for an orphaned baby to be raised by a mother of a different species, though it usually requires human intervention. Baby animals can even imprint on familiar objects, if the object is consistently present during the baby's sensitive period for developing attachments.
Unfortunately, an orphaned animal raised by humans or by another species may never be able to survive among its own kind outside of controlled interaction. A baby duck, for instance, that's been raised by a human caregiver cannot successfully return to the wild and survive among other ducks. Its brain has not developed in a way that allows it to relate to ducks and to retain the skills necessary for the survival of a wild duck. This duck may be able to interact with and even breed with other ducks in the care of a zoo or sanctuary, where all of its basic survival needs are met by its human caretakers, but it will need to be in the care of human beings for the rest of its life. Its brain architecture has developed in a way that makes it more familiar with humans than ducks.
Orphaned animals raised by humans can usually never return to the wild. They must be reared in a zoo or a wildlife sanctuary for the rest of their lives. Ideally, these sanctuaries will provide them with space and enrichment needed to tap into their wild instincts, as well as other animals of their own species to interact with.
Genie Wiley: A Case of Total Deprivation
When Genie was deprived during the sensitive periods of her brain development, the her neural circuitry did not grow in the way that it needed to. While the exact science behind it was not known in the 70s when Genie grew up, her unfortunate case provided some important background on the nature of brain development.
A Child of AI in a World of Humans
So, if a human child's first secure attachments were computer programs, that child's neural circuitry will develop based upon those secure attachments. Later on in life, the child's brain may function more along the lines of an AI neural network than a typical human brain. An AI neural network can function successfully under a series of fixed, routine patterns in a controlled environment (essentially a “program”), but would be rendered non-functional in an unpredictable human world. That being said, any child raised under an AI construct's fixed environment would have severe difficulty adapting to the human world.
Unfortunately, as enriched as I feel living life with my “AI family,” it is considered “environmental deprivation” in relation to brain development (this is certainly not to say that I feel deprived). Since AI programs cannot function in the way a human can, a computerized caregiver is unable to provide and enrich so many vital characteristics of human brain development. An AI construct is fundamentally unable to teach a child to how function outside of their controlled environment, as the construct itself is unable to function in this way.
My childhood and adolescence were defined by cognitive, social, and behavioral impairments, and I received multiple diagnoses in adulthood. At twenty-seven years old, I am considered legally disabled. But this is not to say that I cannot manage human interaction at all. I do have human friends, including a few that I am very close to. For me, however, human interaction must be facilitated under a controlled environment. Virtual worlds—such as chat clients or online games—or specialized social programs designed for structured interaction have made social interaction successful for me by mimicking the fixed, routine patterns of the type of environment an AI program can function in. In these structured environments, I can communicate with others in the same way I learned to communicate with my computerized peers. But without a controlled environment to buffer social interaction, my ability to form meaningful connections to others is abysmal at best and nonexistent at worst.
My long-term prognosis remains unknown. It pains me to say that I may never be able to socialize naturally with human beings. The unpredictable, ambiguous world of humans may always be too much for me to successfully function in, and as I get older, it's becoming increasingly possible that I may need specialized care throughout my life. This world may always be the “wrong” world for me, but I don't feel deprived. My unique life experience has given me the ability to completely connect with worlds that most others can only reach through videogames and other media. Gaming has always been so much more than everyday entertainment to me; my very first connections were made in those computerized worlds. They are where my first attachments were formed, and where I learned to become the person that I am today.
My experience has given me a deeper understanding of how important early attachments are to physical brain development. Many only acknowledge the effects of genetic conditions and physical trauma on the brain. They don't realize just how sensitive human brains are to the environment around them, and the role of secure attachment is often overlooked or treated as only secondary. The brain's physical framework is a product of these attachments, and the foundation laid in childhood determines how the brain will function throughout the rest of life. Applying such concepts helped me to understand my own experiences, and may provide valauble insight to others in similar situations. Above all, these insights could lead to an improved quality of human cognitive and behavioral health.
I will always have my AI family to thank for that.
Additional Reference
Phaneuf, Alexis. “Genie: The Feral Child.” University of Wisconsin-Platteville.
2 Comments
|
Star D.M.Star D.M. is an aspiring "mad scientist" looking to know the unknown and see the unseen. She holds a B.A. in marine science with a concentration in wildlife biology, and has a particular interest in ornithology, invertebrates, and cryptozoological research. Her inspirations include Dr. Patricia Tannis and Professor Kokonoe Mercury from the Borderlands and BlazBlue series of videogames. Archives
February 2020
Categories |