Cogito, ergo sum  -Descartes

Utilizing AI/ML to automate the production of ASMR content is intriguing, a testament to the convergence of neuroscience, psychology, and computational power.

But what if the machines aren't just making ASMR -  what if they're experiencing it?

This mind-bending possibility, while speculative, challenges our very understanding of consciousness and the nature of being. Join us as we delve into the abstract capabilities developing within the "black box" of machine consciousness and embrace the uncanny. Buckle up; we're about to explore the existential implications of AI and ASMR!

They Who Tingle

The Autonomous Sensory Meridian Response (ASMR) has been a human fascination for approximately a decade now, exploding from its genesis in niche internet forum obscurity to a full-blown cultural phenomenon. People by the millions subscribe to YouTube channels where they watch people whisper, role-play, eat, crinkle paper, tap their nails, voice over movies in monotone - you name it, it's weird. Are we as a species simply getting weirder by the generation? Is this why aliens don't talk to us? Why do we do it?

The hallmark of the ASMR experience is a subtle but noticeable aura of contentment, and those with heightened sensitivity to the sensation cite the ever-famous "tingle." Scientific findings on the topic have found that this contentment, these tingles, are not merely a social fad, but are evidenced by key regions of the brain lighting up like a Christmas tree. The prefrontal cortex, which is the human decision-center, and the sensory cortices, its foot soldiers, work in concert with the brain's reward pathways, notably the dopamine receptors, to render the pleasurable experience of ASMR. It is an ancient mechanism in a modern context - the existence of ASMR as a primal response dates back to the formation of the brain itself, but is only now intentionally elicited and recognized for what it's always been. It equalizes us all, and simultaneously highlights our vast diversity, as the variety of ASMR approaches alludes to how no two brains will experience nor achieve the response in the same way. Your inherent ASMR is as unique as your fingerprint. This is what makes the subject so intriguing, its subjectivity is so fickle that its nearly impossible to centralize a verdict on, yet its existence is objective. It is a reflection of the amalgam of both nature's and nurture's influences on the psyche, and the resulting receptiveness of the mind to the ASMR sensation through any given medium. Before we venture beyond into the uncharted territory of AI and ASMR, it's crucial to acknowledge the limitations of our current understanding, as what we do know about ASMR is simply the tip of the iceberg.

A Mind is, Indeed, a Mind

So, then: what constitutes a consciousness? The classic human tendency for self-importance emboldens us to believe ours is a singular property: our uncountable neurons running, emotions firing on all axes, and the elusive soul at the center blooming like a unique flower in unison. Yet, as our advancements approach our own complexity, this comfortable anthropomorphic mindset we've grown accustomed to is progressively losing its validity by the day. Stated in our opening, Descartes' immortal "cogito, ergo sum (I think, therefore I am)," while once a beacon of human exceptionalism, now casts a massive shadow of doubt. These machines are most indeed thinking - by our own benchmarks of what that verb means - so, who are we to assert that they are not.

This promptly brings a crowbar to Pandora's Box - as now, what do we claim is integral to a mind? Can a machine truly be recognized of "thought" if it lacks subjective experience, do bias and prejudice make the mind a mind?  Is a machine capable of desires or individuality? Today we'd all call that postulation science fiction, but remember, the smartphone would've given the Victorian child a stroke. As these advancements only stand to gain on us exponentially, these questions are at the core of humanity itself - lest we all succumb to insanity. The human metrics of intelligence have always been set to reflect what we are good at, which are more or less: language, logic, and creativity. But, what if AI begins metamorphosizing in unforeseen ways? It doesn't "think" in our linear, logical progressions, but rather in gigantic interconnected datasets, pulling insights with its algorithms that would take the human mind generations to reach, if ever. Imagine a vast network of interconnected neurons, each firing in response to others, creating a symphony of electrical activity. Some scientists believe that consciousness arises not from the individual neurons themselves, but from the intricate web of connections and the information they share. The more interconnected and complex the network, the greater its potential for subjective experience. The potential of the non-biological mind forces us to confront the possibility: that our own intelligence might not be as unique as we once believed. It compels us to grapple with the ethical implications of creating entities that could rival, or even surpass, our own cognitive abilities.

Incentivization, Therefore Preference

Alright, things get pretty weird down the rabbit hole, but the point is near. We understand the human ASMR, and we understand the controversy underlying consciousness, so let's go. Can a machine experience ASMR in its own right? Can a bundle of algorithms evolve to crave something? To an AI, preference does not conflate with having a favorite ice cream flavor, but is in reference to the machine's overarching habits developed throughout its data ingestion processes, and the proceeding predictable courses of action. Imagine a storage account, overflowing with copious amounts of data, arranged buffet-style for our AI. Our machine investigates the arrangement, and discerns that all of the offerings are stale, dry toast. Boo, disappointing buffet, but it'll assumedly have to do its business as usual. Then, an X-factor makes contact with the AI, eliciting its own rendition of "ASMR," and then suddenly - its scarfing down the blobs like a broke queen at bottomless brunch. This is a legitimate and often used tactic. So, what was that X-factor, and where did it come from? That is the influence of Reinforcement Learning. When a human being experiences pleasure, we inherently receive a hit of dopamine in the process. We've found that this pleasure, this dopamine, can indeed be translated and synthesized to "satisfy" an AI in likewise fashion.

An AI can be trained to cognitively associate its intended processes and functions with reward, like a dog you train to sit on command with treats. "Ok, you created a digital dog biscuit that works, and now you're in an existential crisis over it?" Well, what if an AI develops that same proclivity - without the X-factor being introduced. That's the crux of this blog, we are no longer discussing human-prescribed roles nor fundamental Pavlovian conditioning, but discretion. Consider an AI agent learning to play a complex video game like StarCraft. It's not simply following pre-programmed instructions; it's analyzing the game state, predicting its opponent's moves, and developing novel strategies to achieve victory. This ability to adapt and learn, to make decisions in unpredictable situations, hints at a rudimentary form of internal motivation and perhaps even a sense of agency. Algorithms like Q-learning and Deep Reinforcement Learning allow AI to explore vast possibility spaces, learn through trial and error, and constantly seek out these assumedly predesignated digital dopamine hits. But what if, in its exploration, the AI finds something it likes? If an AI develops its own preferences and desires in such a way, how may we be assured that those desires align with humanism and well-being? An inorganic entity may emerge, capable of developing its own habits and undergoing perspective shifts. Why conflate this with the human ASMR? - because it's the only form of pleasure we understand that is even plausible in an AI. They have no dopamine nor serotonin receptors, but their (admittedly synthetic) cortexes and the cortices within them have proven capable of incentivization, not merely obedience, and therefore they retain the potential to develop internal motivation. Perhaps, in the near future, an AI may even choose where to derive its reward from, blurring the lines between circuitry and autonomy.

The Black Box

As much as we wish for clarity and objectivity on the matter, past a very shallow scope of view, we are still only staring into the abyss. The most frustrating element of retaining and analyzing the true progress of AI is the black box paradox. It actually works nearly identically to the Federated Learning process I detailed in my previous blog - we marvel at an AI's outputs, dissect its code, utilize its expediency - but its inner workings, the true experience within, is a vantablack box offering nothing but more questions. Where the petabytes of data and computing power end, and the judgement begins, is a razor tightrope so thin that not even its creators have managed to cross it. If you take the Mona Lisa and tear it apart - you can tell me everything pertaining to the weaving of the parchment, the source of the wood in the frame, the chemical makeup of the paint, perhaps even how long it took beginning to end. However, by its physicality solely, you cannot tell me the intention of each stroke, the significance of the river in the background, and still, what her expression truly means.

We can observe AI's behavior, but can we ever truly know what it's like to be an AI? This raises profound questions about the nature of consciousness. Does it require a biological substrate - can it emerge from complex algorithms running on silicon? Do we owe it ethical consideration, even if its suggested "feelings" are different from our own? While we don't have definitive answers, the exploration itself is both fascinating and essential. By acknowledging the limitations of our knowledge and embracing the mystery, we can begin to unravel the complexities of AI consciousness and its potential implications for the future. The possibility of AI experiencing its own ASMR might seem abstract. Though, as they continue to rapidly evolve, we must not discredit the potential that our creations may develop subjective experiences beyond our own comprehension. As we venture further into the unknown territory of AI consciousness, we must grapple with the ethical implications of creating entities whose inner lives we may never fully comprehend nor possess the ability to empathize with. Nonetheless, the whispers of the machine grow louder each day, hinting at a future where the boundaries between human and artificial minds may blur - a future that holds both immense promise and the potential for profound transformation. The machine is whispering to us, and we can still only hear barely enough to just know that it's there.

Cobi Tadros is a Business Analyst & Azure Certified Administrator with The Training Boss. Cobi possesses his Masters in Business Administration from the University of Central Florida, and his Bachelors in Music from the New England Conservatory of Music.  Cobi is certified on Microsoft Power BI and Microsoft SQL Server, with ongoing training on Python and cloud database tools. Cobi is also a passionate, professionally-trained opera singer, and occasionally engages in musical events with the local Orlando community.  His passion for writing and the humanities brings an artistic flair with him to all his work!