Can the "ought" be derived from the "is"?

Recently, I had the privilege to speak one on one, with Swami Sarvapriyananda on the topic of AI Ethics, and also present our work in this area. We presented two of our approaches: one is called "Computational Transcendence" that is aimed at making AI agents act more responsibly, and the second called "Theory of Being" that we argue would be the next evolutionary step in AI-- from "agents" to "beings". We shall not go into the details of our work in this post, and instead, focus on the philosophical talking points that I took back from our discussion. 

Firstly, my thanks to Oxford Dharma Centre and its dynamic founder and volunteers who made this happen, and also for the rest of the members of our initiatives on exploring dharmic foundations for a post AI future (not mentioning specific names here, in case there are any privacy considerations).

*~*~*~*~*

The theme of our discussion can be described by the question: "Can the `ought' be derived from the `is'?". 

Scientific inquiry has predominantly focused on understanding "what is" rather than resolving "what ought to be". The "is-ness" of physical reality is presumed to be impersonal and indifferent to what ought to be the case. And all questions involving AI acting ethically and responsibly seems to be squarely rooted in the realm of "ought". 

The question of ethics is sought to be explained and resolved on the basis of different hermeneutic schools, each of which, identify some fundamental attribute that ought to be pivot for our ethical actions. For instance, utilitarianism is the ethical view that an action is right if it produces the greatest overall happiness or least suffering for the most people, while consequentialism is the broader idea that the moral value of an action depends mainly on its results rather than on the act itself. Deontology takes a different approach, arguing that actions are right or wrong according to expected behaviour in a role-- involving requisite duties, rules, or principles, regardless of the consequences. Virtue ethics in contrast, focuses not on rules or outcomes but on the "character" of the agent-- asking what a good or virtuous person would do. Similarly, Reciprocity is the principle of mutual fairness and exchange, often expressed as treating others as you would like to be treated. Other important ethical views include ethics of care, which emphasises empathy, relationships, and responsibility; divine commandments, which grounds morality in God’s commands; and so on. 

In all of the above, the idea of ethics is grounded in some form of "ought", which makes all schools of ethics as some form of ideology. Indeed, all human social systems are sought to be explained within some ideological framework or the other. 

*~*~*~*~*

But somehow, the worldview in which we were brought up in, we never saw ethics as a matter of ideology-- it was always something deeper and something more fundamental, and rooted in the very fabric of how nature functions. 

Even today, there are several thought leaders who endorse medieval thinkers like Thomas Hobbes, who, in his work Leviathan, describes the "state of nature" (life without government or law) as a "war of all against all." For Hobbes, humans are driven by a constant, relentless desire for power and self-preservation, and are inherently competitive and violent. In this view, morality and ethics are not innate but is a social contract; we agree to act "ethically" only because we fear the chaos of the state of nature and the punishment of a sovereign authority (the state/law) that keeps us in check.

From the worldview that I have been brought up in, this is as alien as it can get. We were taught to regard ethics and morality as something divine-- not in the form of divine commandments that we ought to follow or be liable for punishment-- but as something that represents a refined form of being. A king for example, is evaluated not by his power, but by how well he performs his "Raja dharma", which in our English-medium schools was translated as "duties of a king". Similarly, we learnt that families are run based on "Grihasta dharma" which was then translated as "duties of a householder". 

It is hence not really surprising that the term "dharma" is seen as a confused, overloaded term, and is variously translated as "righteousness", "duty", "divine law", even "religion", and what not-- making our worldview just another form of not so well defined, normative system. 

Before I come back to dharma that forms the grounding for our notion of ethics (and just about everything else), let me also take up one more question: Is ethics a purely objective construct? In other words, can I act ethically in a system where I bear no consequences of my actions? 

For instance, a robot that has been sent to kill a human being (of which, there have already been several documented instances) can justify its actions based on any of the above schools of thought. It can argue that killing that person was necessary for upholding "greater good" (utilitarianism), or by killing the said human, it has performed its expected role properly (deontics), or killing of that person was its way of upholding the virtue of justice, or whatever. 

But there is a different between a robot offering such a justification, and a human who had committed this activity, offering a justification. In the latter case, the human has a stake and is subject to judgment and maybe prosecution. A robot has no such stake. Hence, the "ethical" action of the robot is quite different from that of a human. 

Also, while a robot knows what it is supposed to do, it does not really know why it is doing this activity. For instance, suppose a robot sent to kill a person were to see that the person is already injured and is unable to offer resistance, or if the person is offering to surrender-- it does not change its behaviour, unless it has been explicitly programmed to watch out for these signals and change behaviour accordingly. It does not encounter a "change of heart" suddenly upon seeing the human side of the adversary. Similarly, one cannot use a strategy like "appealing to conscience" of a rampaging robot, to make it stop its destructive activities. 

If we claim that we have resolved "ethics" without the involvement of the subject or their conscience, I'm not sure we have really resolved the question. 

*~*~*~*~*

The emerging post-AI society has reached such a quagmire, where the existential dilemmas are so subtle that people can actually be convinced that a rampaging autonomous robot is acting ethically, and all the damage it has caused in this process, is actually justifiable!

What we are seeing now is the saturation point for machine hermeneutics-- which for the past couple of centuries, has been the dominant interpretive framework worldwide, and which is considered the only framework whose explanations can be considered scientific. In this hermeneutic framework, the universe is seen as an impersonal automaton, grounded in material reality, and driven by the laws of physics, where life somehow emerges out of a large number of random interactions, and in turn gives rise to social constructs like ethics, morality, and consciousness. 

Indian thought (at least most schools of Indian thought) has rejected this idea of material grounding for life and consciousness, since at least 5000 years. In Indian thought, the fundamental substrate of all reality is a triad-- sat-chit-ananda (loosely translated as existence-consciousness-bliss). This is not three separate entities, but one entity that is the ground of all three. The grounding of all physical (material) reality-- the "is" that is investigated by science, as well as the grounding of all humanist concerns including our sense of self, ethics, morality, freedom, etc. and the grounding of all "Axiology"-- that is, our sense of beauty, aesthetics, art, poetry, etc.-- is this same fundamental reality sat-chit-ananda that is the entire universe. 

The dynamics of the existential universe that is grounded in this reality, is itself characterised by different states of being. Each state of being where the system settles down is called its "dharma"-- that represents a sustainable (resilient) state. Each resilient state in turn has different extents of capability (artha) and creative autonomy (kama). (For our purposes, we will not consider the fourth element of Purushartha, that is moksha). 

We can see that while non-living ecosystems (like the surface of Mars) can have different sustainable states, it still represents a lower form of "dharma" as it cannot support life (the chit and ananda realms of reality). A better state of being would be a resilient state where capability for life (including wealth, power, etc.) and creative autonomy are also maximised. This would form the basis for a king to administer his kingdom, or for a householder to run his family.

So yes, our sense of "ought" was indeed derived from the "is". 

*~*~*~*~*

As part of our conversation, Swamiji also narrated an anecdote about Maa Sharada, where she once admonishes a disciple who throws away the broom after sweeping the floor. She asks him to "treat the broom with dignity"! 

While at first, it may appear as someone being overtly sentimental about the "dignity" of inanimate things, this kind of thinking stems from the hermeneutic that treats humanistic constructs like dignity and ethics to be grounded in reality, rather than ideology. While today's India itself has lost much of this kind of approach to life, it can still be seen in other Asian societies. In Japanese culture for example, many practices reflect a deep respect for inanimate objects and physical spaces, treating them as worthy of dignity, care, and even a kind of spiritual presence. This can be seen in traditions such as Shinto, which honours the sacredness of nature and the idea that spirits, or kami, may dwell in places, objects, and natural features. Everyday customs like removing shoes before entering a home, carefully maintaining gardens, wrapping tools with respect, and cleaning public spaces all express the belief that environment shapes behaviour and that order is not merely imposed from outside but cultivated through attention and reverence. This attitude gives objects and spaces a natural sense of civility, as though harmony is embedded in the world itself and should be preserved through disciplined, thoughtful conduct.

Models of Intelligence

Today, whenever we speak of AI, we mean a specific model of AI based on large language models and diffusion, both of which, undergo foundational training in an approach called autoregression. In today's discourse, whenever a researcher asks a "cutting edge" question about AI, they are just asking a question about how to tweak the current model of AI. For instance, whenever researchers ask questions about "AI Ethics", most of them at least, are just asking about LLM Ethics in the form of guardrails, constitution, etc. 

But the idea of intelligence itself (and as a consequence, ethics, morality, etc.) has fascinated researchers over several thousands of years now. And even in the last 70 years or so, many philosophical questions about intelligence have been asked and sought to be explained in precise, mathematical terms. 

In this post, let us have a look at the different ways in which "intelligence" can be understood and modelled. 

At the outset, I should push aside theories like IQ scores, from this post. If someone actually believes that intelligence can be meaningfully represented as a single-dimensional, numerical value, and we can rank people and populations based on this value; then they are not very intelligent, and I hesitate to engage with them. I'm just amazed by how popular is this whole IQ business still.

*~*~*~*~*

The term "Artificial Intelligence" was born in the summer of 1956, at what is now called the Dartmouth Conference. Organized by a young researcher named John McCarthy, the gathering brought together brilliant minds eager to prove that thinking machines were no longer just the stuff of science fiction. McCarthy had a branding problem: he needed a catchy banner to unite researchers scattered across cybernetics, psychology, and early computer science. Other researchers like Herbert Simon and Allen Newell pushed for a much more bureaucratic, academic-sounding alternative-- "Complex Information Processing."

Had they won the debate, today we might be talking about the rapid rise of "CIP" systems and worrying about whether complex information processors are taking our jobs. Instead, McCarthy’s catchy term "Artificial Intelligence" captured popular imagination, transforming a dry engineering pursuit into a thrilling quest to decode the very essence of the mind, capturing the public's imagination and setting the stage for decades of grand dreams, wild hype, and sci-fi obsession.

As a result of this, there are several models of intelligence that have been proposed, each of which, have lead to their own kinds of breakthroughs. Let us have a look at some of these models. 

Intelligence as observable behaviour 

Alan Turing who is credited with creating the mathematical basis for computation, was one of the first to also propose an "empirical" definition of intelligence, in his 1950 paper, "Computing Machinery and Intelligence". He proposed that instead of getting bogged down in unanswerable philosophical debates about intelligence or asking whether machines can truly "think," we should define intelligence entirely through observable behaviour. If a machine acts, converses, and reacts indistinguishably from a human, then for all practical and empirical purposes, it is intelligent. This pragmatic approach makes intuitive sense because even among humans, internal mental states are inherently unobservable in others-- we infer intelligence solely through external outputs like speech and action. There is a related concept that was first used in political science, and now also in computer science. This is called "duck typing": "If we see something that walks like a duck, waddles like a duck, and quacks like a duck, then it is probably a duck." Just as duck typing cares only about an object's methods rather than its internal class inheritance, Turing’s test judges a system purely by its functional interface with the world. Hence, if someone or something acts intelligently, then it is probably intelligent.

However, this behavioural approach has faced pushback from several detractors. Philosophers like John Searle famously countered this with his "Chinese Room argument". Imagine a (non-Chinese) clerk who is stuck in an office where all transactions happen only in Chinese. The clerk is given a set of rules by which he should respond to different sequences of symbols. And suppose the clerk is able to do that. Then, it appears to the outside world that the clerk now understands Chinese. But the clerk is merely following a set of rules. Similarly, while a computer can flawlessly manipulate symbols to simulate fluent conversation, it is not actually understanding what it is doing-- it is merely following some rules. 

Philosopher Peter Norvig, with his "pigeon argument" argues that if aviators were to think of building flying machines in the way Turing thought of building thinking machines, then their goal would be to build a machine "that looks like a pigeon, flies like a pigeon and fools other pigeons into thinking that it is a pigeon." Which is absurd. Our flying machines (airplanes) do not look like birds-- yet they fly. Similarly, we can agree that animals are also intelligent-- yet they cannot carry out a conversation with humans and fool humans into thinking that they are also human.

Yet others like Peter Wegner argue that the Turing Test encourages deceptive superficiality-- rewarding programs that master the art of parlour tricks and linguistic bluffing rather than true reasoning, common sense, or adaptability. For instance, how do you play chess with two grandmasters at once, and fool both of them into thinking that you are a grandmaster? The answer is to get them to play against one another-- through us (that is, see what the first player plays, and make the same move to the second player, and reflect the second player's move, back to the first player). While this trickery can still be argued to be a "manipulative" kind of intelligence, this still leaves open the more fundamental question-- how did the grandmasters get their intelligence?

Intelligence as Imitation: 

One of the key markers of intelligence that we see in children (including cubs of many animals) is their ability to imitate their surrounding. Children often copy their parents and elders in their mannerisms and action and further on, even in thought and interpretation. 

Imitation requires an enormous amount of pattern matching and reasoning capabilities. A "completely imitative agent" of a human for example, should be able to mimic a human in all our capabilities-- in actions, thoughts, and even in ethics, morals and spiritual grounding. Hence, an agent that can maximally imitate the universe, can in fact, imitate the intelligence in the universe, and hence, can be called intelligent as well. 

Current day LLMs are a result of this paradigm. The core model in an LLM is called "autoregression" where a corpus of text is its own training data. An LLM foundation model does not require a separate tagged dataset and a supervisor to teach the LLM. The LLM randomly masks words in the text and tries to predict it based on the surrounding context. In essence, it is trying to imitate the corpus, and if the corpus is large enough and contains enough amount of intelligent output, the imitating agent also appears intelligent. 

This is the reason why many now believe that machines have defeated the Turing test. Today's AI makes their output pretty much indistinguishable from human output. (How much of this blog post was generated by AI-- can you guess?) 

But of course, the main issue with this interpretation of intelligence is that, imitation requires an intelligence environment (corpus) for it to imitate-- and the question of where did the intelligence in the environment (corpus) come from, still remains open. 

Intelligence as conceptual modelling: 

This is yet another approach propounded by the conceptual modelling community. To explain this, we first look at the parable of the "hungry cat". 

A man had a pet cat at home. Everyday when he returned home from work, his cat would be quite hungry. It leaps at his legs and starts demanding to be fed. It keeps coming between his feet, making him trip. With great difficulty, he reaches the refrigerator to take out some cat food. But due to the relentless demands of the cat, which is jumping and leaping at the food, he invariably ends up dropping the food on the floor, which the cat eats and goes away. 

The exasperated man goes to the animal psychologist to ask why isn't the cat learning to sit by its food bowl and wait for him to feed it. It can see that he intends to feed it. 

The animal psychologist replies that the cat is indeed learning something. It has learned to "predict the next token"-- if it keeps leaping at your feet and demanding food, it has predicted that, food will fall from above! 

This incidentally, is exactly how present day LLMs work as well-- by predicting the next token (word) given a sequence of words. 

What the cat was not able to do, was to build a conceptual model based on understanding latent elements like intention, and an understanding of what the person intends to do. 

While conceptual modelling involves a lot of formal logic and symbolic reasoning, building the conceptual model itself is beyond this. As the saying goes, "We can only communicate truth using logic-- we cannot discover truth with logic." A good model helps us perform symbolic reasoning and inference-- not the other way around. With just formal logic and symbolic reasoning, we cannot automatically build good models. 

There is no exact "science" of conceptual modelling. We all build conceptual models at different extents of explainability and accuracy. A good conceptual model forms a good abstraction-- in that, it captures the "essence" of what is being observed, which is typically a latent entity. Capturing the essence goes beyond just the ability to predict the next token. A good conceptual model offers a good explanation and understanding of the situation, rather than just matching patterns. 

Yet another saying goes: "All models are wrong, but some are useful". For instance, Newtonian physics is a remarkable conceptual model that captures the essence of physical dynamics of the universe. But even Newtonian physics is "wrong" in the sense that, it breaks down when we try to apply this model in the sub-atomic or inter-galactic scale. 

Regardless of that, our ability to build conceptual models that goes beyond just predicting the next observation, and offering an explanation and understanding of underlying principles, is an indicator of intelligence. 

Intelligence as agency: 

This line of argument says that if we observe every entity that we consider intelligent-- humans, animals and even plants-- all have one thing in common. They are all "living" beings and have some element of "free will". 

All naturally intelligent beings are autonomous. They don't wait to be prompted or commanded. They act on their own, and pursue their self interest. 

This line of argument gave rise to the field of "Intelligent agents" that tried to ask what is meant by autonomy-- if autonomy is the marker for intelligence. Being able to "break down" a prompt into an action sequence and execute them (which is what present day agentic LLMs do) is an artefact of autonomous action, but is not autonomy in itself. 

I often give this example to illustrate the above. Sometimes when walking in the darkness outside, we mistake a coil of rope for a snake, and freak out. But the question is, so what if it is a snake? What is the problem? Is it that the snake can kill us? Well, even a rope can kill us-- if we fashion it into a noose and hang ourselves from it. The thing is-- a snake can autonomously kill us! We don't have to do anything for the snake to kill us. While we need to keep "prompting" the rope into forming a noose and hanging us from it, before we can get killed by it..

So this line of argument says that, unless machines can have their own self interest (that is, they will "want" to take over the world or at least want to fight for their "rights"), they are not exactly intelligent. They are adaptive, predictive pattern matching automata-- rather than intelligent agents that exercise their autonomy. 

Intelligence as embodied cognition: 

In this line of argument, intelligence is seen as having a purpose-- to help protect the body. All intelligent beings we saw, act autonomously. But we see that acting autonomously does not mean they act in an arbitrary or random fashion. 

All autonomous actions of living beings fundamentally stem from their quest for survival. And hence, as this line of argument goes, intelligence is tightly tied to our bodies. According to this line of argument, purely software forms of intelligence, including LLMs, are not really intelligent, because they have no purpose of their own. Rodney Brooks, one of they key researcher argues that "Elephants don't play chess" as a critique at purely software based approaches to intelligence like chess playing programs. 

*~*~*~*~*

Many of these lines of argument have given rise to key technologies and insights that are still critical to support present day AI. For instance, the entire field of compilers came up because we wanted to formalise language and thought. Similarly, the embodied cognition argument gave rise to present day robotics that are vastly better and autonomous, than earlier forms of robots. 

In our quest for asking cutting-edge questions, I do hope we do not assume AI means LLMs and intelligence means autocomplete, or explainability means symbolic logic. The real problems are much deeper. 

The limitations of machine hermeneutics

What we observe is not nature itself, but nature exposed to our method of inquiry --Werner Heisenberg 

Consider this scenario: A client approaches a lawyer and says, "My friend has cheated me!". In response to which, the lawyer asks two questions: "Why did your friend cheat you?" and "How did your friend cheat you?" 

Clearly, they are two different questions and requires two different answers. The first one asks about motives and intentions behind the cheating, and the second one asks about the mechanism of the act. 

Now consider this scenario: A school student approaches a scientist and asks two questions: "Why does a lunar eclipse occur?" and "How does a lunar eclipse occur?" The answer to both of these questions is the same! 

In our scientific models of physical reality, there is no difference between "why" and "how"! This is because, in our models of physical reality, there is no such thing called "motive", "intention", and such elements of "free will" as a fundamental element of reality. 

This is what I call "machine hermeneutics". 

For the last few centuries, we have been adopting a specific hermeneutic or "way of inquiry" to understand and interpret reality. In this approach, we consider the entire universe to be an impersonal automaton powered by "energy" and whose dynamics are characterized by causality that began with the big bang. Everything else-- including life and its associated constructs like free-will, intentions, motives, etc. are all considered to be complex emergent properties of these impersonal causal dynamics. 

Machine hermeneutics began in medieval Europe in the 18-19th century, and was greatly influenced by the industrial revolution. In this mode of inquiry, the study of reality and the design of machines go hand in hand. Machine hermeneutics replaced earlier models of reality that were overtly anthropomorphic. For instance, a physical theory that says that a lunar eclipse happens when the "moon covers itself with shame" or something like that, is (justifiably) considered unscientific, and discarded. 

Focusing only on the mechanics of physical phenomena has clearly brought us great benefits. We can how build machines inspired by such phenomena and extend our capabilities and reach. Which is precisely what we have been doing for the last few centuries. 

We have built more and more complex machines, so much so that the machines are now starting to replicate the ultimate frontier of human faculty-- our brains and intelligence. 

Our machines today can replicate the mechanics of a number of activities that we thought were driven by the "human spirit"-- like creating artwork, writing poems, translating text, carrying on a conversation with another human, etc. Machines today can also autonomously take decisions and act on them. 

The mechanics of human intelligence seems to have been unraveled! Or have they? Can we finally close the gap between why humans are intelligent and how humans are intelligent? 

*~*~*~*~*

Consider this case from 2016. where the police in Dallas used a robot to kill a suspect who was engaged with a shootout with them. Or the use of autonomous killing drones in the Ukraine war. There are videos that show (like this one) drones pursuing soldiers and taking them down, with the soldiers pretty much helpless to do anything against them. The more interesting (and disturbing) fact is in some of the pictures, the body language of the soldiers show that they have clearly given up-- and yet, the drone takes them down. 

There is this axiom that drives human soldiers to kill on the battlefield-- "I do not want to kill someone with whom I have no personal enmity over, but if I do not kill the enemy soldier, he might kill me!" 

In other words, humans (and all other living beings) hold their lives to be precious and resort to killing other living beings mostly as a means to preserve their own life (except when driven by ideology). An autonomous drone on the other hand, has no fear for its own life, and has no empathy for the other person. It is just programmed to kill, and cannot tell when the adversary is no longer a threat, and need not be taken down. 

The AI in the drone only knows how to make the kill, and not why it is killing! 

Before our lives were taken over by machines, the why question was all pervasive and played a central role in our daily activities. It was also clearly distinct from the how question. We often had solutions that solely involved addressing the why question rather than the how question. 

Suppose that we are back in the medieval times before the advent of automobiles, and we owned a horse carriage pulled by two horses. For the last few weeks after we bought a new horse, we notice that our journeys in the horse carriage have been very bumpy and uncomfortable because the two horses that are pulling the carriage are not synchronizing with one another. Before getting to ask how to make them coordinate with one another, we first ask, why are they acting in an uncoordinated manner. It may turn out that one of the horses that was newly acquired, is feeling insecure, or is in need of our attention and care. In such a case, all we may need to do is to spend more quality time with the horse in question and get it better acquainted with us. 

Now imagine the same situation above in the present day, when we buy a new car, and find that it is not running well. We can be certainly sure that the reason for this would be something more mechanical-- like a faulty engine tune up or a loose connection somewhere. 

But imagine AI driven cars in the not-so-distant future that learns about your preferences and customizes itself over time. Now, spending "quality time" with the car, suddenly becomes a thing! As cars get more intelligent, we may need them to understand not just how to drive us somewhere, but also why we're going somewhere and what it means! For instance, an autonomous car taking us to the airport to receive our family members may have different considerations to address from the same car taking us to a hospital near the airport in response to a medical emergency. 

For too long, we have been thinking of reality as the ultimate machine. Even theories like Quantum Mechanics, only doubt the deterministic nature of mechanical causality (at least in the Copenhagen interpretation that interprets the wave function as the uncertainty in the knowledge of the observer), but do not give primacy to free-will itself. We will need a very different way of interpreting the universe as machines become an integral part of our lives and well-being.