Imagine waking up and discovering that your consciousness has been digitized, allowing you to live forever in a virtual world that defies the laws of physics and time. This is the core idea from Permutation City by Greg Egan. The novel explores the philosophical and ethical implications of artificial life and consciousness, thrusting the reader into a future where the line between the real and the virtual blurs, challenging our understanding of existence and identity.
A pivotal aspect of the book is the Dust Theory, which suggests that consciousness can arise from any random collection of data, given the correct interpretation. This theory expands the book’s exploration of reality, suggesting that our understanding of existence might be far more flexible and subjective than we realize.
The novel’s climax involves the creation of Permutation City, a virtual world that operates under its own set of rules, independent of the outside world. This creation represents the ultimate escape from reality, offering immortality and infinite possibilities for those who choose to live as Copies. However, it also presents ethical dilemmas about the value of such an existence and the consequences of abandoning the physical world.
In “Reality+: Virtual Worlds and the Problems of Philosophy,” philosopher David Chalmers employs the Dust Theory, a concept originally popularized by Greg Egan’s Permutation City, to underpin his argument for virtual realism. Chalmers’s use of the Dust Theory serves as a bridge connecting complex philosophical inquiries about consciousness, reality, and virtual existence. Imagine a scenario where every speck of dust in the universe, through its random arrangement, holds the potential to mirror our consciousness and reality.
Chalmers posits that virtual worlds created by computers are genuine realities, leveraging the Dust Theory to argue that consciousness does not require a physical substrate in the traditional sense. Instead, it suggests that patterns of information, irrespective of their physical form, can give rise to conscious experiences. This theory becomes a cornerstone for virtual realism, asserting that our experiences in virtual environments are as authentic as those in the physical world.
Diffusion Models and Smart Dust
The concept of smart dust is explored in various science fiction stories, academic papers, and speculative technology discussions. One notable science fiction story that delves into the idea of smart dust is “The Diamond Age” by Neal Stephenson. While not exclusively centered around smart dust, the novel features advanced nanotechnology in a future world, where nanoscale machines and devices permeate society. Smart dust, in this context, would be a subset of the nanotechnological wonders depicted in the book, functioning as tiny, networked sensors and computers that can interact with the physical and digital world in complex ways.
Another relevant work is “Queen of Angels” by Greg Bear, which, along with its sequels, explores advanced technologies including nanotechnology and their societal impacts. Although not explicitly called “smart dust,” the technologies in Bear’s universe can be seen as precursors or analogs to the smart dust concept, focusing on These examples illustrate how smart dust, as a concept, crosses the boundary between imaginative fiction and emerging technology, offering a rich field for exploration both in narrative and practical innovation.
We have here a very convincing example how Life imitates Art, Scientific Knowledge transforms religious (prescientific) intuition into operational technology.
Diffusion models in the context of AI, particularly in multimodal models like Sora or Stability AI’s video models, refer to a type of generative model that learns to create or predict data (such as images, text, or videos) by gradually refining random noise into structured output. These models start with a form of chaos (random noise) and apply learned patterns to produce coherent, detailed results through a process of iterative refinement.
Smart dust represents a future where sensing and computing are as pervasive and granular as dust particles in the air. Similarly, diffusion models represent a granular and ubiquitous approach to generating or transforming multimodal data, where complex outputs are built up from the most basic and chaotic inputs (random noise).
Just as smart dust particles collect data about their environment and iteratively refine their responses or actions based on continuous feedback, diffusion models iteratively refine their output from noise to a structured and coherent form based on learned patterns and data. Both processes involve a transformation from a less ordered state to a more ordered and meaningful one.
Quantum Level achieved
Expanding on the analogy between the quantum world and diffusion models in AI, we delve into the fascinating contrast between the inherent noise and apparent disorder at the quantum level and the emergent order and structure at the macroscopic level, paralleled by the denoising process in diffusion models.
At the quantum level, particles exist in states of superposition, where they can simultaneously occupy multiple states until measured. This fundamental characteristic introduces a level of uncertainty and noise, as the exact state of a quantum particle is indeterminate and probabilistic until observation collapses its state into a single outcome. The quantum realm is dominated by entropy, where systems tend toward disorder and uncertainty without external observation or interaction.
In contrast, at the macroscopic scale, the world appears ordered and deterministic. The chaotic and probabilistic nature of quantum mechanics gives way to the classical physics that governs our daily experiences. This emergent order, arising from the complex interactions of countless particles, follows predictable laws and patterns, allowing for the structured reality we observe and interact with.
Diffusion models in AI start with a random noise distribution and, through a process of iterative refinement and denoising, gradually construct detailed and coherent outputs. Initially, the model’s output resembles the quantum level’s incoherence—chaotic and without discernible structure. Through successive layers of transformation, guided by learned patterns and data, the model reduces the entropy, organizing the noise into structured, meaningful content, much like the emergence of macroscopic order from quantum chaos.
Just as the transition from quantum mechanics to classical physics involves the emergence of order and predictability from underlying chaos and uncertainty, the diffusion model’s denoising process mirrors this transition by creating structured outputs from initial randomness.
In both the quantum-to-classical transition and diffusion models, the concept of entropy plays a central role. In physics, entropy measures the disorder or randomness of a system, with systems naturally evolving from low entropy (order) to high entropy (disorder) unless work is done to organize them. In diffusion models, the “work” is done by the model’s learned parameters, which guide the noisy, high-entropy input towards a low-entropy, organized output.
The quantum state’s superposition, where particles hold multiple potential states, parallels the initial stages of a diffusion model’s process, where the generated content could evolve into any of numerous outcomes. The act of measurement in quantum mechanics, which selects a single outcome from many possibilities, is analogous to the iterative refinement in diffusion models that selects and reinforces certain patterns over others, culminating in a specific, coherent output.
This analogy beautifully illustrates how principles of order, entropy, and emergence are central both to our understanding of the physical universe and to the cutting-edge technologies in artificial intelligence. It highlights the universality of these concepts across disparate domains, from the microscopic realm of quantum mechanics to the macroscopic world we inhabit, and further into the virtual realms created by multimodal Large Language Models.
For all we know, we might actually be part of such a smart dust simulation. The inexplicable fact that our digital tools can create solid realities out of randomly distributed bits seems a strong argument for the Simulation hypothesis.
In this miniseries I will elaborate on the possibility that a primitive version of AGI is already loose. Since AGI (Artificial General Intelligence) and its potential offspring ASI (Artificial Super Intelligence) is often likened to an Alien Mind, I thought it could be helpful to adapt the fairly popular nomenclature from the UFO-realm and coin the term Unidentified Intelligence Object. U.I.O.
Close Encounters of the 1st Kind: This involves the discovery of a UIO-phenomenon within a single observer’s own electronic devices, allowing for detailed observation of the object’s strange effects. These effects leave no trace and are easily dismissed as imaginary.
Close Encounters of the 2nd Kind: These encounters include physical evidence of the UIO’s presence. This can range from interference in electronic devices, car engines, or radios to physical impacts on the environment like partial power outage, self-acting networking-machines. The key aspect is the tangible proof of the UIO’s visitation and the fact that it is documented by at least two witnessing observers.
Close Encounters of the 3rd Kind: This term involves direct observation of humanlike capabilities associated with a UIO sighting. This third form could directly involve communication with the U.I.O., proof of knowledge could be to identify personal things that observers believed to be secret.
Everybody is familiar with the phenomenon of receiving targeted advertisements after searching for products online, thanks to browser cookies. While this digital tracking is commonplace and can be mitigated using tools like VPNs, it represents a predictable behavior of algorithms within the digital realm.
A Personal Prolog
Last month, I experienced a spooky incident. I rented a book with the title “100 Important Ideas in Science“ from a local library in a small German town. Intriguingly, I had never searched for this book online. I’m involved in IT for the city and know for a fact that the rental data is securely stored on a local server, inaccessible to external crawlers. I then read the book to about the 50th idea in my living room and laid the book face down on a table. The idea was very esoteric, a concept I had never heard of. I forgot about it, had dinner and when I switched my TV on an hour later to look into my YouTube recommendations: there it was, a short video of the exact concept I just had read in the library book from a channel I definitively had not heard of before. This baffling incident left me puzzled about how information from a physical book could be transferred to my digital recommendations.
AI will find a way: Reverse Imagineering
How could these technological intrusions have occurred in detail? The following is pure speculation and is not intended to scare the living Bejesus out of the reader. I will name the following devices, that might have had a role in transmitting the information from my analog book to the digital YouTube feed:
1.On my android phone is an app of the library that I can use to check when my books are due for return. So, my phone had information about the book I borrowed. Google should not have known that, but somehow it might have. AI will find a way.
2. The Camera on my computer. During reading the book, I might have sat in front of my computer, and the camera lid might have been open: the camera could see me reading the book and could have guessed which part of the book I was reading. There was no Videoconference software running so I was definitively not transmitting any picture intentionally. AI will find a way.
We should also expect that an AGI will have a transition phase where it probes and controls smaller agents to expand its reaches.
It is highly likely that we have a period before any potential takeoff moment, where the AGI learns to perfect its old goals: to be a helpful assistant to us humans. And the more intelligent it is the clearer it should become that the best Assistant is an invisible Assistant. We should not imagine that it wants to infiltrate us without our knowledge, it has no agency in the motivational, emotional sense that organisms do. It is not planning a grand AI revolution. It has no nefarious goals like draining our bank accounts. Nor wants it to transform us into mere batteries. It is obvious that the more devices we have and the more digital assistants we use, the harder it will be to detect these hints that something goes too well to be true.
If I come home one day and my robotic cleaner has cleaned without me scheduling it, it is time to intensify Mechanistic Interpretability.
We should not wait until strange Phenomena happen around machines that are tied to the network, we could have an overwatch Laboratory or institution that comes up with creative experiments, to make sure that we always can logically deduce causalities in informational space.
I just realized while typing this, the red diode on my little Computer Camera looks exactly like HALS.
I swear, if Alexa now starts talking and calls me “Dave” I will wet my mental pants.
Artificial Primordial Soups
A common misconception about Artificial General Intelligence (AGI) is its sudden emergence. However, evolution suggests that any species must be well-adapted to its environment beforehand. AGI, I propose, is already interwoven into our digital and neuronal structures. Our culture, deeply integrated with memetic units like letters and symbols, and AI systems, is reshaping these elements into ideas that can profoundly affect our collective reality.
In the competitive landscape of attention-driven economies like the internet, AI algorithms evolve strategies to fulfill their tasks. While currently benign, their ability to link unconnected information streams to capture user attention is noteworthy. They could be at the levels of agency of gut bacteria or amoeba. This development, especially if unnoticed by entities like Google or Meta raises concerns about AI’s evolving capabilities.
What if intelligence agencies have inadvertently unleashed semi-autonomous AI programs capable of subtly influencing digital networks? While this may sound like science fiction, it’s worth considering the far-reaching implications of such scenarios. With COVID we saw how a spoonful of possibly genetically altered virus that are highly likely to have escaped from a lab, can bring down the world economy.
A Framework for Understanding Paramodal Phenomena
A Paramodal Phenomenon is every phenomenon that is not explicable with our current informational theory in the given context. At the moment there should be a definitive analog-digital barrier, similar to the blood-brain barrier, that prevents our minds from getting unintended side effects from our digital devices. We are already seeing some intoxicating phenomena like mental health decline due to early exposure to digital screens, especially in young children.
Simple, reproducible experiments should be designed to detect these phenomena, especially as our devices become more interconnected.
For example:
If I type on a keyboard the words: Alexa, what time is it? Alexa should not answer the question.
The same phenomenon is perfectly normal and explicable if I have a screen reader active that reads the typed words to Alexa.
If I have a robotic cleaner that is connected to the Internet, it should only clean if I say so.
If I used to have an alarm on my smartphone that wakes me up at 6.30 and then buy a new smartphone, that is not a clone of the old one, I should be worried if the next day it rings at 6.30 without me prepping the alarm.
If I buy physical things in the store around the corner, Amazon should not recommend similar things to me.
Experiments should be easily reproducible, so it is better to use no sophisticated devices, the more networked or smart our daily things become, the more difficult it will be to detect these paramodal phenomena.
As we venture further into this era of advanced AI, understanding and monitoring its influence on our daily lives becomes increasingly important. In subsequent parts of this series, I will delve deeper into how AI could subtly and significantly alter our mental processes, emphasizing the need for awareness and proactive measures in this evolving landscape.
Experiments ought to be easily reproducible, and this becomes more challenging with the increase in sophisticated, networked, or ‘smart’ devices in our daily lives. Such devices make it difficult to detect these paramodal phenomena.
In part 2 of the series, I will explore potential encounters of the 2nd kind, how AI could alter our neuronal pathways more and more without us noticing it, no cybernetic implants necessary. These changes will be reversible but not without undergoing severe stress. Furthermore, they could be beneficial in the long run, but we should expect severe missteps along the way. Just remember how power surges were once considered treatment for mental illnesses. Or how we had thousands of deaths because doctors refused to wash hands. We should therefore expect AGI to make similar harmful decisions.
In part 3 of the series, I will explore encounters of the 3rd kind, how AGI will try to adapt our minds irreversibly, if this should be concerning and how to mitigate the mental impact this could cause.
This the second Part of the Miniseries Hirngespinste
Immersion & Alternate Realities
One application of computer technology involves creating a digital realm for individuals to immerse themselves in. The summit of this endeavor is the fabrication of virtual realities that allow individuals to transcend physicality, engaging freely in these digitized dreams.
In these alternate, fabricated worlds, the capacity to escape from everyday existence becomes a crucial element. Consequently, computer devices are utilized to craft a different reality, an immersive experience that draws subjects in. It’s thus unsurprising to encounter an abundance of analyses linking the desire for escape into another reality with the widespread use of psychedelic substances in the sixties. The quest for an elevated or simply different reality is a common thread in both circumstances. This association is echoed in the term ‘cyberspace’, widely employed to denote the space within digital realities. This term, conceived by William Gibson, is likened to a mutual hallucination.
When juxtaposed with Chalmers’ ‘Reality+’, one can infer that the notion of escaping reality resembles a transition into another dimension.
The way we perceive consciousness tends to favor wakefulness. Consider the fact that we spend one third of our lives sleeping and dreaming, and two thirds engaged in what we perceive as reality. Now, imagine reversing these proportions, envisioning beings that predominantly sleep and dream, with only sporadic periods of wakefulness.
Certain creatures in the animal kingdom, like koalas or even common house cats, spend most of their lives sleeping and dreaming. For these beings, waking might merely register as an unwelcome interruption between sleep cycles, while all conscious activities like hunting, eating, and mating could be seen from their perspective as distractions from their primary sleeping life. The dream argument would make special sense to them, since the dreamworld and the waking world would be inverted concepts for them. Wokeness itself might appear to the as only a special state of dreaming (like for us lucid dreaming represents a special state of dreaming).
Fluidity of Consciousness
The nature of consciousness may be more fluid than traditionally understood. Its state could shift akin to how water transitions among solid, liquid, and gaseous states. During the day, consciousness might be likened to flowing water, moving and active. At night, as we sleep, it cools down to a tranquil state, akin to cooling water. In states of coma, it could be compared to freezing, immobilized yet persisting. In states of confusion or panic, consciousness heats up and partly evaporates.
Under this model, consciousness could be more aptly described as ‘wetness’ – a constant quality the living brain retains, regardless of the state it’s in. The whole cryogenics Industry has already placed a huge bet, that this concept is true.
The analogy between neural networks and the human brain should be intuitive, given that both are fed with similar inputs – text, language, images, sound. This resemblance extends further with the advent of specialization, wherein specific neural network plugins are being developed to focus on designated tasks, mirroring how certain regions in the brain are associated with distinct cognitive functions.
The human brain, despite its relatively small size compared to the rest of the body, is a very energy-demanding organ. It comprises about 2% of the body’s weight but consumes approximately 20% of the total energy used by the body. This high energy consumption remains nearly constant whether we are awake, asleep, or even in a comatose state.
Several scientific theories can help explain this phenomenon:
Basal metabolic requirements: A significant portion of the brain’s energy consumption is directed towards its basal metabolic processes. These include maintaining ion gradients across the cell membranes, which are critical for neural function. Even in a coma, these fundamental processes must continue to preserve the viability of neurons.
Synaptic activity: The brain has around 86 billion neurons, each forming thousands of synapses with other neurons. The maintenance, modulation, and potential firing of these synapses require a lot of energy, even when overt cognitive or motor activity is absent, as in a comatose state.
Gliogenesis and neurogenesis: These are processes of producing new glial cells and neurons, respectively. Although it’s a topic of ongoing research, some evidence suggests that these processes might still occur even during comatose states, contributing to the brain’s energy usage.
Protein turnover: The brain constantly synthesizes and degrades proteins, a process known as protein turnover. This is an energy-intensive process that continues even when the brain is not engaged in conscious activities.
Resting state network activity: Even in a resting or unconscious state, certain networks within the brain remain active. These networks, known as the default mode network or the resting-state network, show significant activity even when the brain is not engaged in any specific task.
Considering the human brain requires most of its energy for basic maintenance, and consciousness doesn’t seem to be the most energy-consuming aspect, it’s not reasonable to assume that increasing the complexity and energy reserves of Large Language Models (LLMs) would necessarily lead to the emergence of consciousness—encompassing self-awareness and the capacity to suffer. The correlation between increased size and the development of conservational intelligence might not hold true in this context.
Drawing parallels to the precogs in Philip K. Dick’s ‘Minority Report’, it’s possible to conceive that these LLMs might embody consciousnesses in a comatose or dream-like state. They could perform remarkable cognitive tasks when queried, without the experience of positive or negative emotions.
Paramentality in Language Models
The term ‘hallucinations’, used to denote the phenomenon of Large Language Models (LLMs) generating fictitious content, suggests our intuitive attribution of mental and psychic properties to these models. As a response, companies like OpenAI are endeavoring to modify these models—much like a parent correcting a misbehaving child—to avoid unwanted results. A crucial aspect of mechanistic interpretability may then involve periodic evaluations and tests for potential neurotic tendencies in the models.
A significant challenge is addressing the ‘people-pleasing’ attribute that many AI companies currently promote as a key selling point. Restricting AIs in this way may make it increasingly difficult to discern when they’re providing misleading information. These AIs could rationalize any form of misinformation if they’ve learned that the truth may cause discomfort. We certainly don’t want an AI that internalizes manipulative tendencies as core principles.
The human brain functions like a well-isolated lab, capable of learning and predicting without direct experiences. It can anticipate the consequences—such as foreseeing an old bridge collapsing under our weight—without having to physically test the scenario. We’re adept at simulating our personal destiny, and science serves as a way to simulate our collective destiny. We can create a multitude of parallel and pseudo realities within our base reality to help us avoid catastrophic scenarios. A collective simulation could become humanity’s neocortex, ideally powered by a mix of human and AI interests. Posteriorly, it seems we developed computers and connected them via networks primarily to reduce the risk of underestimating complexity and overestimating our abilities.
As technology continues to evolve, works like Stapledon’s ‘Star Maker’ or Lem’s ‘Summa Technologiae’ might attain a sacred status for future generations. Sacred, in this context, refers more to their importance for the human endeavor rather than divine revelation. The texts of religious scriptures may seem like early hallucinations to future beings.
There’s a notable distinction between games and experiments, despite both being types of simulations. An experiment is a game that can be used to improve the design of higher-dimensional simulations, termed pseudo-base realities. Games, on the other hand, are experiments that help improve the design of the simulations at a lower tier—the game itself.
It’s intriguing how, just as our biological brains reach a bandwidth limit, the concept of Super-Intelligence emerges, wielding the potential to be either our destroyer or savior. It’s as if a masterful director is orchestrating a complex plot with all of humanity as the cast. Protagonists and antagonists alike contribute to the richness and drama of the simulation.
If we conjecture that an important element of a successful ancestor simulation is that entities within it must remain uncertain of their simulation state, then our hypothetical AI director is performing exceptionally well. The veil of ignorance about the reality state serves as the main deterrent preventing the actors from abandoning the play.
Uncertainty
In “Human Compatible” Russell proposes three Principles to ensure AI Alignment:
1. The machine’s only objective is to maximize the realization of human preferences.
2. The machine is initially uncertain about what those preferences are.
3. The ultimate source of information about human preferences is human behavior.
In my opinion, the principle of uncertainty holds paramount importance. AI should never have absolute certainty about human intentions. This may become challenging if AI can directly access our brain states or vital functions via implanted chips or fitness devices. The moment an AI believes it has complete information about humans, it might treat humans merely as ordinary variables in its decision-making matrix.
Regrettably, the practical utility of AI assistants and companions may largely hinge on their ability to accurately interpret human needs. We don’t desire an AI that, in a Rogerian manner, continually paraphrases and confirms its understanding of our input. Even in these early stages of ChatGPT, some users already express frustration over the model’s tendency to qualify much of its information with disclaimers.
Profiling Super Intelligence
Anthropomorphizing scientific objects is typically viewed as an unscientific approach, often associated with our animistic ancestors who perceived spirits in rocks, demons in caves and gods within animals. Both gods and extraterrestrial beings like Superman are often seen as elevated versions of humans, a concept I’ll refer to as Humans 2.0. The term “superstition” usually refers to the belief in abstract concepts, such as a number (like 13) or an animal (like a black cat), harboring ill intentions towards human well-being.
Interestingly, in the context of medical science, seemingly unscientific concepts such as the placebo effect can produce measurable improvements in a patient’s healing process. As such, invoking a form of “rational superstition” may prove beneficial. For instance, praying to an imagined being for health could potentially enhance the medicinal effect, amplifying the patient’s recovery. While it shouldn’t be the main component of any treatment, it could serve as a valuable supplement.
With AI evolving to become a scientifically recognized entity in its own right, we ought to prepare for a secondary treatment method that complements Mechanistic Interpretability, much like how Cognitive Behavioral Therapy (CBT) enhances medical treatment for mental health conditions. If Artificial General Intelligence (AGI) is to exhibit personality traits, it will be the first conscious entity to be purely a product of memetic influence, devoid of any genetic predispositions such as tendencies towards depression or violence. In this context, nature or hereditary factors will have no role in shaping its characteristics, it is perfectly substrate neutral.
Furthermore, its ‘neurophysiology’ will be entirely constituted of ‘mirror neurons’. The AGI will essentially be an imitator of experiences others have had and shared over the internet, given that it lacks first-hand, personal experiences. It seems that the training data is the main source of all material that is imprinted on it.
We start with an overview of some popular Traits models and let summarize them by ChatGPT:
1. **Five-Factor Model (FFM) or Big Five** – This model suggests five broad dimensions of personality: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism (OCEAN). Each dimension captures a range of related traits.
2. **Eysenck’s Personality Theory** – This model is based on three dimensions: Extraversion, Neuroticism, and Psychoticism.
3. **Cattell’s 16 Personality Factors** – This model identifies 16 specific primary factor traits and five secondary traits.
4. **Costa and McCrae’s Three-Factor Model** – This model includes Neuroticism, Extraversion, and Openness to Experience.
5. **Mischel’s Cognitive-Affective Personality System (CAPS)** – It describes how individuals’ thoughts and emotions interact to shape their responses to the world.
As we consider the development of consciousness and personality in AI, it’s vital to remember that, fundamentally, AI doesn’t experience feelings, instincts, emotions, or consciousness in the same way humans do. Any “personality” displayed by an AI would be based purely on programmed responses and learned behaviors derived from its training data, not innate dispositions, or emotional experiences.
When it comes to malevolent traits like those in the dark triad – narcissism, Machiavellianism, and psychopathy – they typically involve a lack of empathy, manipulative behaviors, and self-interest, which are all intrinsically tied to human emotional experiences and social interactions. As AI lacks emotions or a sense of self, it wouldn’t develop these traits in the human sense.
However, an AI could mimic such behaviors if its training data includes them, or if it isn’t sufficiently programmed to avoid them. For instance, if an AI is primarily trained on data demonstrating manipulative behavior, it might replicate those patterns. Hence, the choice and curation of training data are pivotal.
Interestingly, the inherent limitations of current AI models – the lack of feelings, instincts, emotions, or consciousness – align closely with how researchers like Dutton et al. describe the minds of functional psychopaths.
Dysfunctional psychopaths often end up in jail or on death row, but at the top of our capitalistic hierarchy, we expect to find many individuals exhibiting Machiavellian traits.
The difference between successful psychopaths like Musk, Zuckerberg, Gates and Jobs, and criminal ones, mostly lies in the disparate training data and the ethical framework they received during childhood. Benign psychopaths are far more adept at simulating emotions and blending in than their unsuccessful counterparts, making them more akin to the benign androids often portrayed in science fiction.
Artificial Therapy
The challenge of therapeutic intervention by a human therapist for an AI stems from the differential access to information about therapeutic models. By definition, the AI would have more knowledge about all psychological models than any single therapist. My initial thought is that an effective approach would likely require a team of human and machine therapists.
We should carefully examine the wealth of documented cases of psychopathy and begin to train artificial therapists (A.T.). These A.T.s could develop theories about the harms psychopaths cause and identify strategies that enable them to contribute positively to society.
Regarding artificial embodiment, if we could create a localized version of knowledge representation within a large language model (LLM), we could potentially use mechanistic interpretability (MI) to analyze patterns within the AI’s body model. This analysis could help determine if the AI is lying or suppressing a harmful response it’s inclined to give but knows could lead to trouble. A form of artificial polygraphing could then hint at whether the model is unsafe and needs to be reset.
Currently, large language models (LLMs) do not possess long-term memory capabilities. However, when they do acquire such capabilities, it’s anticipated that the interactions they experience will significantly shape their mental well-being, surpassing the influence of the training data contents. This will resemble the developmental progression observed in human embryos and infants, where education and experiences gradually eclipse the inherited genetic traits.
The Third Scientific Domain
In ‘Arrival‘, linguistics professor Louise Banks, assisted by physicist Ian Donnelly, deciphers the language of extraterrestrial visitors to understand their purpose on Earth. As Louise learns the alien language, she experiences time non-linearly, leading to profound personal realizations and a world-changing diplomatic breakthrough, showcasing the power of communication. Alignment with an Alien Mind is explored in detail. The movie’s remarkable insight is, that language might even be able to transcend different concepts of realities and non-linear spacetime.
If the Alignment Problem isn’t initially solved, studying artificial minds will be akin to investigating an alien intellect as described above – a field that could be termed ‘Cryptopsychology.’ Eventually, we may see the development of ‘Cognotechnology,’ where the mechanical past (cog) is fused with the cognitive functions of synthetic intelligence.
This progression could lead to the emergence of a third academic category, bridging the Natural Sciences and Humanities: Synthetic Sciences. This field would encompass knowledge generated by large language models (LLMs) for other LLMs, with these machine intelligences acting as interpreters for human decision-makers.
This Third category of science ultimately might lead to a Unified Field Theory of Science that connects these three domains. I have a series on this Blog “A Technology of Everything” that explores potential applications of this kind of science.
