All cultures develop machines and industries in their own image and hence an effort to build such machinery through an authentic Indic and Dharmic perspective should be our aim.
Continued From Part 1
The human ideals of Christian Thought and its Corresponding Critique on AI
Before we question the possible utility of AI, we need to decide on the ultimate ideals of human life. When we build machines to reduce the physical or mental labour of people, to what use would people put that flexibility in life? In the western thought process, apart from the theories of Greek philosophers, the Christian religion had a profound influence. This religious thought nurtured a fear about AI, that the creation of artificial mind as a competition to the very creation of life by God is a signal of the arrogance of man which would ultimately cause his downfall. In addition, Christianity also provides guidance for improving AI. This influence can be discussed in three aspects.
Causality: Firstly, originating from Judaism, is the theory that the Creator God of the universe is distinct and separate from the Creation, and that there is no divinity present in any object of the universe. The same idea is present in Christianity and Islam. In this theory, amongst the various objects of creation, only man received the capacity for intellectual consciousness (soul), while the other beings such as plants and animals are appointed to be of use to man. The ultimate ideal of man is in worshipping the Creator, and in having faith over His authority on the Creation. Every phenomenon in the universe has a cause, and when these causes are arranged as a chain, the causal chain should ultimately lead to a Primal Cause which is the Creator. Thus, the intellect that can perform causal investigation (Kāraṇa Vichāraṇa) needs to ultimately reach the state of worship of the Creator. If it stops at any earlier state, these religions would consider that to be an inferior intellect (Alpa Buddhi). Thus, Causality (Kāraṇatva) has a foundational basis within these religions, even as it is also considered an important area of study within AI. In the western system of mathematics, causality manifests in the process of developing a rational proof that connects a professed theorem to the primary postulates (Axioms). In the modern science of physics, the germination of the physical laws in the universe, their eventual transformation, the growth of stars and planet, thereafter the evolution of various life-forms on earth, should all be mapped on to a straight historical line of causality. When the causal chain is connected to each other, that should lead to the Primal Cause (whether we call this as God or not is immaterial). This emphasis on causality is also the reason why AI is seen overarchingly by many people as a tool to “predict the future”. Various algorithms can learn how to make such predictions based on data, but it is important to separate mere correlation from causation. From this perspective of causality, AI can be improved by analyzing its results, and producing causal interpretations of the generated phenomena to prior accepted mathematical or ethical principles, for example, those deriving from the Bible. However, this emphasis on Linear Causality (Rēkhīya Kāraṇatva) is not present in all cultures. Indeed, in Indian philosophical thought, this was repudiated. It may be futile to even try to predict certain real-world phenomena, irrespective of how much data is accumulated. Equivalently, it may be inadequate to correct such predictions based on whichever prior assumptions.
Regulation by Law: Secondly, according to Christianity, since humans don’t have divinity within themselves, they require information delivered as Word (Śabda) to unravel the secrets of the universe. This revelation of information from God happens through the means of Prophets, in whose speech the information is revealed in human language. These utterances (Uvācha) are sufficient for the achievement of the ultimate human ideals, whichever they may be. Even if we put aside the special status of the Prophets, there Is an underlying belief in the capacity of human language to encode this information within a finite set of words. The question of whether this is even possible did not arise in these civilizations. When the completeness or perfection of language is not questioned, the truth value of any proposition in that language is associated with two sides (true or false). It is expected that these two sides will cover all the possibilities of the universe. Such a binary logic (Dvipaksha Tarka) has been condemned in Indian philosophical thought. The Hindu opinion is that logical examination (Tarka Pariśīlana) in any limited language would be equivalent to slavery to that language. Consequently, Indian logic considered that any logical proposition doesn’t have two sides, but multiple sides. The most famous of these multi-sided logics (Bahupaksha Tarka) is the four-sided Chatushkōṭi, which considers four sides to a proposition (A, Not A, Both A and Not A, Neither A nor Not A). Because of this difference in logic, placing overarching authority in Word (Śabda) or in a book of instructions is repudiated in Indian culture. Similarly, in Chinese philosophical traditions, multifaceted logics such as I-Ching represent nuanced notions of uncertainty. In contrast, in the religious practices of Christians or Muslims, it is believed that the Word can be effortlessly and unambiguously translated into any human language and that converting people by spreading that Word is a possibility. This would be considered incorrect in Indian philosophical thought. According to the Christian perspective, artificial intelligence can be evaluated based on whether its usage is in accordance with the principles given by God through the Word of Prophets, and codified in Holy Books such as the Bible. Such an analysis can take place for both the phenomena generated by AI and for the interactive use of AI by humans. In the secular perspective, the holy book such as the Bible is replaced by another book, such as the constitution, but the same type of juridical analysis is needed. However, from the perspective of Indian philosophical thought, such a limited juridical analysis in language would not be effective.
Individual Rights: Thirdly, Christianity states that each human received an individual soul from God. There is no interconnectedness or relationship between the souls of different people, or between their souls and nature. Due to this view on isolated existence, there is an emphasis on the individuality of each human, and the consequent individual human rights. When an AI is seen with the same isolationary perspective, it will also be seen as an individual being with individual desires. It does not have any deep connectedness with humans or other life forms in nature. In a world which is run by the dictates of such an AI, it becomes imperative to provide dedicated individual rights to humans or other life-forms. These digital rights of humans in relation to AI should be codified in books of law. However, this isolated existence of individual souls is repudiated in Indian philosophical thought. Chinese traditions such as Daoism and Confucianism are also opposed to such anthropocentric perspective and reject isolationary individualism. Hence, their perspectives on AI would also differ substantially.
To sum up, according to Christian ideals, AI cannot behave arbitrarily without any rules. For responsible AI, criteria such as transparency and human oversight can be judged on the basis of Christian ideals. The actions of digital life can be examined with respect to the principles laid down in the Holy Book. It is not particularly problematic if such an examination is done through a centralized authority. The other criteria of responsible AI, such as robustness, fairness and human well-being, can also be decided on the basis of Christian ideals. The equivalent decisions for the ethical improvement of AI, which are taken by other cultures, will naturally be different.
Ārya Prajñā: Artificial Intelligence as an instrument towards achieving Purushārthas
The human ideals in Indian culture are known as Purushārthas. These are four in number: Kāma, Artha, Dharma and Mōksha. Kāma refers to bodily pleasures, whereas Artha refers to mental pleasures, such as money and possessions. Ultimately, money is just a mental abstraction, which if agreed upon by a whole society, provides the means for man to achieve various bodily pleasures. It is possible to use AI to achieve Artha and Kāma. This is a fully justified pursuit of Purushārthas. However, in this effort, the other two Purushārthas should not be forgotten.
Dharma refers to the principles required for the healthy life of man, or any being (Jīva). The same is known in the modern era as sustainability (Susthiratva, Nitya Pushṭitva). In the Dharmic perspective, it is obligatory to examine whether by the use of AI, if nature is incurring any harm, or if any human relations are incurring any harm. The application of AI in opposition to the natural Dharma of any being should not be done. Every human has a personalized set of principles (Dharma), based on his or her birth, and on the spatio-temporal context of the country and community (Dēśakālasamaya). Every being in nature has a specialized individualized Dharma. When the world overall runs in accordance with Dharma, everybody would receive pleasure and happiness. The cosmic cycle of Ṛta is the basis for the Dharma of any being (Jīva). On the basis of Ṛtadharma, rainfall occurs, crops grow, and the cycle of time marches onward. When every man respects this Ṛtadharma, and respects the sacredness (Pavitrata) in nature, and performs the various auspicious acts at the right time (Puṇyakārya), that will be the fulfilment of individualized Dharma (Svadharma). It is possible to use AI as an instrument to the praxis of Dharma (Dharmācharaṇa).
Mōksha refers to the liberation (Vimukti) from bodily suffering (Aihikaduḥkha). As explained earlier, this is possible only through the sipping of the nectar of knowledge (Jñānāmṛta). Different Dharmic traditions suggest different meditative paths for this. If AI is developed into an assistive tool towards this, it should be a tool for exercising the physical and mental activities of man and thus nurture their health. The use of AI should promote memory capacity (Jñāpaka Śakti), capacity to hold a thought (Dhāraṇa Śakti), the unmoving capacity to contemplate (Dhyāna Paṭutva) and so on. In opposition to these, the AI should not destroy the native human mental faculties, by transforming man to use mechanical tools as supporting rods and inculcating various forms of mental addictions.
It is possible to use the Indian ethical perspective as a contrast to the above mentioned western ethical values, in driving the development of responsible AI.
Interconnectedness of Sacrifice (Yajñabandha) instead of Causality (Kāraṇatva): According to Dharmic thought, nature is self-created. It is not created by any external Creator. A cause cannot be attributed to every phenomenon in the universe. While the gross phenomena are caused by subtle phenomena, the subtle phenomena cannot all be traced into a causal chain leading to a primal cause such as God (Bhagavān). Such an extreme causality is entertained only in the Dvaita tradition of the Dharmic schools. The Dharmic thought attributes divinity to all the natural elements: fire, water, stars, sun and the moon, whose creative forces are all deeply interconnected with each other. It is said that these Dēvatas support each other, and while doing so, drive the wheel of time forward. Such voluntary sharing and helping without desiring anything in return is termed Yajña. Even humans are expected to realize this divinity within their own inner beings and lead their lives like a Yajña. It is not just humans, the endless Yajña of this universe has all the living and non-living beings as participants. Therefore, artificial intelligence will also be a partner in this Yajña. According to the Dharmic perspective, improving an AI is not seen by reducing them into individual isolated entitites, but by developing harmony and interoperability between them. These artificial programs should inform each other, evaluate each other’s results and together build a clear pool of knowledge (Satva Jñanasampada). The AIs should also be an assistance to the Dharma of different living beings.
Prosperity of refined flowing language (Āryabhāshāsaṃvṛddhi) instead of regulation by law: Refined flowing culture (Āryasamskṛti) refers to the above-explained realization of Yajña relationships and interconnectedness in everyday life. The word “Āryam” is derived from the sense of “flow”. The Sanskrit root Ṛ (Gatikaraṇé: to flow) is transformed by the suffix “nyat” according to Pāṇini’s grammatical rule “Ṛhalōnyat”, and produces the word “Āryam”. This Āryan culture refers to developing irrigation networks from river basins, turning the land green through sustainable farming, and thereby, ensuring happiness to all beings. In this culture, the very embodiment of knowledge is a great river, Sarasvatī. In Indian history, the cultivated Indo-Gangetic river plains were known as Āryāvarta. It is immaterial that foolish western Indologists attributed racist interpretations to this word. But Indian culture is essentially refined flowing culture (Āryasaṃskṛti). In this perspective, all tools for artificial intelligence are representations of the knowledge goddess Sarasvatī and distributaries of the eternal river. These canals of knowledge-water should not be polluted. To prevent such pollution, naturally we need regulation by law. But even more importantly, the interconnectedness of these canals and the sacredness of water (knowledge) needs to be realized. Repositories of knowledge should arise as water catchments and lakes. Bountiful gardens must grow around them and cater to diverse living beings. When life is seen as a societal celebration, and AI becomes a part of it, all the humans would be experts in its use. This expertise would percolate into their own language, enabling even complex thoughts and phenomena to be expressed as words in that language. Such a language is Āryabhāsha, which will be used for communication between people, as well as for building and interacting with AI. When so many eyes and tongues are actively working in this field, there will not be any pollution in the AI. That prosperity of refined flowing language (Āryabhāshāsaṃvṛddhi) will be the safeguard for the Dharma of everyone.
Environmental awareness instead of Isolation: In the Dharmic perspective, all beings of nature have their existence interconnected with each other. The Vēdas term such interrelated existence as Indrajāla. Buddhists visualize this as how the dew drops on a spider-web inter-reflect the light from each other. Āyurvēda says that for a man to be healthy, the whole society has to be healthy, and for a society to be healthy, the entire natural ecosystem has to be healthy. The same health should be present also in our mental activities. We can decide on the individual rights of users of AI. However, even more extensively, we need to understand the responsibilities and Dharma of the users. This Praxis of Dharma (Dharmācharaṇa) cannot be driven by law. The right type of training (Śikshaṇa) is needed, along with a lot of patience. If AI is built as an instrument to achieve the Purushārthas, then different Dharmic traditions can mould these AI tools into their rituals, activities and meditative practices. With extensive praxis of Dharma, the whole society would flourish in all the wings. This holistic environmental awareness will be built on the basis of the cosmic cycle (Ṛta).
The current tools of AI are often diametrically opposed to the ideal of Āryaprajña: indiscriminate extraction of data, piling them into concentrated repositories, and hiding them from people reflects an extractionary culture that mines mineral wealth from the land, and not a refined flowing culture (Āryasaṃskṛti). They reflect the practices of a desert civilization where water is extracted as a mineral resource and not a land blessed with monsoons. In a refined Ārya society, like water, this data would flow and reach all people. The data will not be collected into a giant cesspool of a repository, where it accumulates various mental prejudices (Manōvikāra) and biases of people. If these rivers of knowledge keep flowing actively, and if the human discriminatory intellect purifies these waters like sunlight and makes them rain all over the land, it would produce cultural artefacts in everyday life. This intelligence (Prajñā) can be instituted into the society at large when data-driven cultural artforms become a part of the mental constituency (Chitta) of people, and make them green (happy) just like the rivers do the gardens and fields. The language of people becomes refined with understanding and data-driven wisdom, ultimately promoting meditative capacity (Dhyānaśīlata). The tools of AI would not bring people’s downfall through mental addictions.
In order to build these Purushārtha-driven ethical values into AI tools, naturally, a lot of effort needs to be expended. The current technological sphere needs to be seen with critical reflection, and new tools need to be created. The education of people in artificial intelligence needs to be provided through their own mother tongues. This entire technology needs to be part of the culture, and our cultural art forms need to be the tools for Āryaprajñā. This can bring the ethical treatises (Dharmaśāstras) and philosophical thought of India into the modern era. However, this responsibility lies entirely with the Indians themselves. Any culture would develop machines and industries in its own image. The effort to build such machinery through an authentic Indic and Dharmic perspective has not yet begun. May this effort reach its fruition soon! Śrīghramēva kāryasiddhirastu!
References for Further Reading:
- This article by Stephen Cave presents a critique on the western philosophical perspectives of intelligence.
- In this article, I give an introduction to the multifaceted Indian logical tradition of Chatuṣkōṭi.
- The text of the Chāndōgya Upanishad, with the aspects of intelligence in the instruction of Bhūma Vidya, is available here with a commentary from the Dvaita tradition.
- The text of the Bammanavagga from the Buddhist Dhammapādas, with an explanation in English, is available here.
- In this prior essay, I described the cosmic cycle of Ṛta and its deep symbolism in Indian
- In this article by Bing Song, she discusses the non-anthropocentric nature of Chinese traditions and how they provide a different perspective on AI, which is somewhat comparable to the Indian perspective presented here.
- Here is the English translation of the Sāṃkhya Sūtras of Kapila by James Ballantyne.