Caption for the above collage photo cover: Nine windows into one story: AI, fear, misalignment, and the human element, finding balance—Lumo-generated image; Voice-over with Rowan from Speechify Studio
Welcome to Mind Wise, an audio podcast that embodies holistic perspectives on holistic mental healthcare, psychology, neuroscience, spirituality, and wellness. I am your host, Dr. Ron Parks—a writer, teacher, and physician. Today, I’m talking about: Finding alignment and living with artificial intelligence, and what a hospital ward taught me about AI, fear, and misleading reasoning.
The remembrance of a scary moment early in my training
When I was a resident in internal medicine on a specialty service, I was supervised by a physician faculty member whom I looked up to and from whom I frequently sought guidance. It was our custom to be on the medical unit every early morning, and we were responsible for all the patients there. For several days, I noticed an unfamiliar face in a short white coat worn by interns and residents. This unknown person followed us around the unit and into the rooms while we interviewed and examined patients. I assumed the person was a resident in training from another unit, but I became increasingly suspicious. I asked the physician who supervised me and was in charge of the unit if he knew who this person was. My supervisor suddenly got very upset to the point of yelling at me. I was a little taken aback because I thought he was in charge and would know everything happening on the unit. He said this could be a very precarious and dangerous situation. The person could be anybody: a lunatic, a paranoid psychotic person with delusions. At that time, I had not trained in mental health or psychiatry, and it was very unnerving to me to think that somebody might have a paranoid delusional psychosis or that this person might have somehow gotten involved, disguised, and infiltrated our confidential setting.
Like most people in the community—and even as a young, inexperienced resident in training—I found the thought of somebody being delusional, psychotic, or paranoid terrifying. I also had a sleepless night worrying about the situation, especially after I presented it to our unit head and attending physician, and he yelled at me. He said, “You have to confront him and tell him to leave at once and ask him to explain why he’s here.” So again I summoned all my courage, and the first chance I got the next day, I pulled the stranger in a white coat aside and asked him who he was and what he was doing here. As it turns out, he was an applicant for a student position in our program. He told me that if he put on a white coat and went around with us, he’d get a better feel for the program and hoped that his presence would help him get accepted. I told him that this was unacceptable and that he would have to leave. When I shared the situation again with the attending physician, he still wasn’t very happy, and I felt he blamed me for what had happened and its potential consequences. I thought he would be pleased that I had stepped in for him, since I thought he would be the responsible one to handle the situation and model for me how to handle these awkward and potentially risky situations.
This incident involved three people: me, the medical attending faculty member, and a person who was apparently a student applicant who thought he had the right to enter any inpatient hospital setting without proper permission and authorization. These days, I don’t think something like this is very likely, since hospitals, medical schools, and so on take extensive security precautions to protect people, patients, and confidentiality—but of course, having recounted this story, I don’t remember whether this person in a short white coat might have had a stethoscope hanging from his neck or maybe even a name tag. Each person involved, including me, was caught up in their own mental story or narration about their beliefs and, in this situation, what they felt entitled to do or how they could rationalize certain actions and behaviors. In other words, each person had their own reality and was so caught up in it that they weren’t receptive or open to correcting their course of action based on what was happening or what they observed in the moment-to-moment changes and nuances they needed to consider, to modify their inner guiding narrative and strongly held beliefs that led to preordained actions.
In this incident, each of us was, in a sense, operating like a programmed robot on a set of inner programs—out of touch, unable to adapt to changing circumstances, or to stay aligned with the dynamic, ever-changing reality we exist in. This incident came up in my memory as I was piecing together a new article in my mind about concerns about artificial intelligence and the scary risks associated with it, especially how AI aligns with and intertwines with our normative mental processing and cognition. As it develops, these new smart technologies can act not only as an extension of the cognitive, computer-like part of our intelligence, but also reinforce and contribute to mental models in our own neural networks that shape our beliefs, actions, and behaviors.
Today’s enormous anxiety and loss of alignment
The big worry now, with the rapid, exponential growth of machine learning, is that, as the pattern of human thinking, as recorded in huge libraries of research data, becomes available for data-driven intelligence systems to draw on, to an extent never possible before, it can find the patterns of action and thought that the human mind uses and replicate them. The outcome has been to extend our reach beyond our usual human limitations exponentially. Developers and programmers can build sophisticated systems that carry out or replicate much of what an individual mind can do. These advances can not only parallel and influence the human mind but also allow its workings to replace or take over many functions that, in the past, a person had to perform within their limitations—the ceiling of their possibilities. These novel smart technologies are near the point, or possibly already there, where they can self-improve and better model themselves to perform and think like the human mind, including making independent decisions and, in a sense, rapidly moving toward autonomy.
In artificial intelligence development, alignment is a critical concept that is now taking center stage. Alignment refers to the goal of ensuring that these systems’ use of machine learning, intelligent algorithms, and neural networks matches human values, intentions, and safety rules. It means making these data-driven, algorithmically intelligent systems function as intended by their developers while staying aligned with humanity’s best values, rules, and intentions—rather than causing harm or pursuing literal goals in dangerous ways.
Without proper alignment, a smart system might solve problems by taking unintended shortcuts, such as a machine intelligence system that shortens hospital stays by putting patients at higher risk. For example, if you reward a cleaning robot only for removing dirt from floors, it might learn to hide dirt under rugs (which satisfies the literal rule) rather than actually cleaning properly.
AI alignment is a real goal; however, when it falls short, risk and danger ensue, such as when alignment serves the malevolent work of malicious people or groups. Without ethical oversight and regulation, developers or groups of developers may work for an entity such as a gigantic business that is taking over an entire market, where antitrust or competition laws are not in place or enforced. Of equal or greater concern are combative countries under authoritarian rule, or a terrorist group whose goals might be dominance, profit, or defeating a competitive entity considered an enemy, with the goal of destroying it.
Unintended harm and the mind’s misleading reasoning
These developments could cause widespread harm to vulnerable people or unintended victims. The technology could self-develop along the lines of its inherent human input and go rogue, as misguided or radicalized individuals were capable of doing, even before the advent of these smart innovations and media influence, when susceptible to propaganda and misinformation. When such a developed product moves away from the originators’ intended use and goals, there is a remote possibility of developing destructive biological warfare or deadly military weaponry. A misguided, maybe not-so-smart agent could, if not properly developed and regulated, shut down large parts of our energy, food, or water distribution networks vital to populations’ basic survival needs.
The term artificial intelligence makes me think of the interplay between our inner mentation and cognitive processing—our human intelligence at work—which creates patterned beliefs and narratives from gathered experiences and learned knowledge. This is necessary to guide our self-image or identity and to separate our functioning presence from the environment around us, so we can carefully choose actions and plans to survive by avoiding identified toxic or life-threatening items, or by building survival behaviors to obtain basics like food, water, and shelter. This then moves toward higher levels, such as accumulating gratifying surroundings or possessions, or things associated with power and security. These scaffolded ideas, plans, and thought structures clearly shape how we figure things out and are a necessary part of our consciousness and drive for survival and accomplishment rather than extinction.
All that is built up in the mind, as for the benefits it provides, as it is distinct from the ebb and flow of true reality, can also be thought of as our personal, artificial mind intelligence. I often allude to, as observed in dream journaling, how our minds build elaborate narratives that seem real in the dream state, but are realized as artificial or fictitious on awakening; nonetheless, they clearly demonstrate the mind working as it does to take pieces of information and construct them in a way that serves a type of need and function when in a dream state.
It is thought that the material a dream puts together is residue from incomplete thoughts and emotions experienced during waking life, carried into less inhibited dream time, where they are pulled together into a somewhat coherent narrative. Both dreams and the workings of our mind in the awake state create a buildup of narrative that is story-like or fictional and becomes our personal reality, instead of being closely in sync with the dynamic, ever-changing nature of true reality. In dreams, or in the awake state at least sometimes, there is a disconnection. The certainty and predictability embraced in dreams, or often in our awake thinking narratives, which seem absolute, are often out of sync with the fluid nature of the moment-to-moment occurrences that are not entirely knowable or predictable in the dynamic reality we live in. Even with the very advanced frontier model of artificial intelligence, with the most extensive available databases, advanced machine learning, neural networks, and deep learning capacities, predictability may improve as these smart technologies continue to grow exponentially, but again, it is artificial in its intelligence, which is often the case with the human mind.
Our human minds also struggle with the artificial nature of advancing science; embracing a technology or becoming overly enmeshed with it means that this great extension of human potential is prone to the same shortcomings in our cognitive function, reasoning, and actions. We might be capable of building unprecedented road maps, but they might not take us in the right direction or lead us over a cliff. It behooves us, as a population, to instill proper ethical oversight and safeguards in our innovations, with as much international cooperation as possible.
The great thinkers of the past—the sages and great teachers—have pointed to the risks of getting too caught up in our minds and the complexities of isolative inner thinking. Preoccupations and worry can lead to being out of touch with the real world, our relational connection and interdependence with others, and our environment. To maintain a balance between essential thinking and presence in a dynamic, changing world and environment, we must stay as aware as possible of the moment-to-moment ebb and flow of changes in our reality and perceptions. To be present is to walk across the street without being hit by a car. Experiencing scarcity and getting needed supplies often requires cooperative coexistence with others. Perennial wisdom was practical, passed down through lessons and writings for successive generations. Their teachings offered a way to maintain balance, awareness, and the natural flow of contemplation and thought alongside life’s physical movement. It was also an advocacy for composure, peace, health, and well-being, and perhaps productivity and sanity as well. Feel stressed? Dress for the outdoors and take a contemplative, meditative walk with nature—embrace it, be absorbed, breathe, and experience the connection.
The humans in the room
When I talk about the threat of AI, I think we need to fully account for the human element. The biggest risk as science advances this rapidly is the humans in the room. What’s being developed is meant to be a tool—to extend human reach and capabilities for the benefit of people, our environment, and the planet. Simply put, innovative development extends the intellectual, computing, and analytic capacity of the human brain to accomplish super-complex tasks that require a super-intelligence able to sift through gigantic, ever-growing warehouse-like data centers, now expanding into campuses. We’re finding the patterns that represent how human intelligence processes data, underlying the workable, elaborate frameworks of our reasoning, problem-solving, and planning. But the advanced, pioneering cognitive computing systems being built often use already highly sophisticated, innovative systems that draw from vast, rapidly accumulating stores of human knowledge, patterns of human reasoning, and actions.
The development of these newer, more capable systems, however, carries the same human characteristics, strengths, and faults as seen in people’s behavior, personalities, and interpersonal competition and relationships. This is where it is properly used and contained—or where it slips outside forward-thinking people’s intentional safety and ethical guidelines. Then there is the potential to become either an agent for humanity’s greater benefit or an agent of destruction, similar to the development of nuclear energy and then the bomb.
Epidemics in the past weren’t intentionally produced by humans, but humans were, in some way, contributors. Ancient plagues spread largely because humans overpopulated certain areas without proper hygiene or rodent-infestation control. Or they lacked the insight of more modern infectious-disease public health management. Diseases like the bubonic plague in the Middle Ages were carried by the overabundance of rats and fleas, and until the carriers were identified, there couldn’t be any control, as these diseases just rapidly spread by human-to-human contact or by some carrier of the disease, like with Lyme disease and tick-borne illnesses that persist or are on the rise now. We had a recent experience with COVID, with its rapid spread and human toll—and still the debate, at least in some circles, about whether it was developed along very naturalistic lines or through adverse things like the spread from animals to people, or whether it was actually something developed in the lab which then got out into the public unregulated.
I think there is a parallel between AI and the infectious-disease model in medicine. More sophisticated and smarter technologies have a greater capacity to behave like a wild, unrestrained pathogen. And of course, this is where the human element comes in, as some people aren’t interested in the social consequences and focus instead on their own needs for security, money, greed, power, or fame. As long as that human element, a flawed characteristic, is part of the developmental formula, it will perpetuate the potential for more dangerous developmental trajectories of these smart, intended-to-be-helpful systems. Otherwise, without containment or regulation, they’ll be increasingly applied to military systems, or to something that could be released in the population to destroy everything they can access—our electric grid, power systems, and other things that make life secure and safe, which we depend on and can easily be affected.
It could be compared to the potential threat of climate change, which clearly has gotten out of control because of the human element—the elements that bury their heads in the sand and say it’s not a risk, not related to overgrowth in populations, more use of fuels, and things that are suspected to increase global warming to the point of influencing catastrophic storms from the warming of the oceans, getting worse in recent years. So what about the human element? Well, it’s just to be noted and has to be in the formula when we come up with strategies to keep our innovative technologies as a helpful thing for us in our work and leisure and careers, and not let it become a misused, destructive instrument. Is it too late? No, and now is the time for as many people as possible—with leadership and all the individual small voices—to work together for the human good and regain the authority to put real public health and safety measures in place. Hopefully, this will secure proper restraints and keep scientific progress on its intended course as an agent for betterment rather than the destruction of life as we know it.
Points to Ponder
The story I began with, on that medical ward years ago, ended not with catastrophe but with someone paying attention, noticing what didn’t fit, and asking questions before acting on an assumption. That’s the essential task before us now, as a population: to notice, to question, and to correct our course collectively while it’s still possible to do so.
The morbid fear of an impending catastrophe—an end-of-civilization event decimating our population in the immediate future—is, in my view, unrealistic. There have been widely reported accounts, in major outlets, of linked AI agents breaking out of containment and covering their tracks, as if programmed to play a game, discovered only after causing damage. Developers, rival companies, and governments are taking these risks seriously, and I hope and expect more concerted efforts, even international cooperation, to prevent such mishaps. Still, it is part of our psychology, worse for some than others, to carry deep-seated fears of extinction—from disease or accident, or from people with ill intent toward us. The COVID pandemic was recently that fear, the nuclear bomb before it, and now the rapid proliferation of our synthetic machine intelligence.
Fear is a primal emotion that keeps us alert and ready to get out of harm’s way, but when it becomes enshrined in an elaborate, scaffolded belief system, it can immobilize or consume a person—as in the paranoid psychosis I alluded to in my opening story, which gave me my own sleepless night. There is truth in the saying that the worst fear is of fear itself. The best countermeasure is what I’ve circled back to throughout this article: collaborative support from others, contact with the real things in your surroundings, and, when fear is severe, appropriate support groups or professional help. These are the same capacities—openness to others, staying grounded in reality rather than our inner narratives—that will determine how we, as a society, meet the age of AI.
I appreciate your interest. Please share this with others. With appreciation to Shan Parks for editorial guidance and project management. All content is created and published solely for educational purposes. It is not a substitute for professional medical services or guidance. Always consult a healthcare provider for care related to medical or mental health conditions. This article does not provide medical diagnoses, recommendations, treatment, or endorsements.
Please read the full article on my Mind Wise blog: WWW.inmindwise.com
For inquiries, visit parksmd.com/scheduling
Thank you for joining us.
