Various applications use AI to only distract human users, not as a means of Tantra (software/meditative framework). But why not? According to the Indian philosophical thought, there is a very lofty ideal for the development of AI.
How do we translate “artificial intelligence” (AI) into Telugu or other Indian languages? In Telugu, the word “Kṛtrimamēdha” is commonly used. Recently, the Sanskrit news program Vārtā on DD News used the phrase “Kṛtrima Buddhimatā” while referring to the recent RAISE summit on Artificial Intelligence, that has been organized by the Indian government. A few others also use the phrase “Kṛtrima Prajñā”. All these words Mēdhā, Buddhī or Prajnā can be used as synonyms to “intelligence” in English. There are several more words in Sanskrit and other Indian languages, with slight variations in meaning between them. Such fine distinctions are not present in European languages. In this essay, however, I would argue that all these translations are unsatisfactory and that an authentic Indic perspective to examine the terminology of artificial intelligence from its foundations is still missing.
Words in English such as “artificial intelligence” have centuries of history and cultural values behind them. Depending on the language we use, the thoughts and impressions in our mind develop a subtle prejudice and tilt our ideals towards the future subconsciously in a subtle manner. However, this future need not be in alignment with Indian ethics or with the Dhārmic cultural values which arose here. Few people have the discrimination to examine this in detail. Therefore, it is necessary to investigate the ideas expressed through the technological terminology at a fine scale, and thereby sharpen our own opinions. If we are able to honestly introspect and interpret our ethical values to technical knowledge, we would be able to offer an authentic perspective to the world and guide the future evolution of the ideas.
Development of Responsible AI: Issues to be considered
AI tools have now moved beyond research labs and have developed into daily instruments that are used by common people. In many industries, government apparatus, health, education and similar spheres, AI is being used regularly to take decisions or to analyze data. Nevertheless, there is no guarantee that the decisions taken thus may not cause any harm to society or to individuals. So Responsible AI (Bādhyatāyuta Kṛtrimamēdhā) arose as a prominent field of study. For example, in the medical field, AI can be used within a clinical diagnosis or within the analysis of the medical history of the patient. Such use is considered responsible AI if it satisfies five criteria:
- Transparency (Pāradarśakatā): The facility to examine why certain decisions are taken
- Oversight (Uparidṛshṭi): The possibility to control the AI by a human expert
- Robustness (Sattā): The AI should behave responsibly even when it was given unexpected inputs or put in unforeseen situations
- Fairness (Maryādā): The AI should behave fairly across all sections of society, and produce impartial results irrespective of racial, sectorial, religious or gender differences. At the same time, it should offer the right type of support to different users based on their context.
- Well-being (Śrēyas): The AI should provide well-being for people and society, and not cause any harm.
All these issues can be discussed in detail depending on the situational context. To ensure that AI acts responsibly on all these criteria is an eternal task. As times change, and as technology evolves, the situational context for judging these criteria will also change. The aptness of the behaviour, or the acceptability of the results produced by AI, need to be judged anew each time. Any AI is not an isolated object, but intimately tied to the overall technological sphere, and is a part of the public infrastructure available to the society. Hence, just like providing roads, ports, electricity or irrigation networks are considered a public responsibility, providing access to the internet, and to the AI hardware and software should be considered as the technological infrastructure of the nation, which is a duty of the government.
When the government sets up such a heavy infrastructure, it needs to also ensure the right mix and balance between various trade-offs: the answerability of the government, the division of powers between different agencies, the educational opportunities for people to improve their mastery with the use of AI, the promotion of commercial enterprises in achieving economic growth while simultaneously setting up the right regulation to prevent exploitation and injustice. Currently, AI expertise is concentrated primarily in the USA and China. The indiscriminate extraction of data in those countries, extraordinary power being concentrated in just a few agencies (whether private companies or government agencies), and the exploitation of AI for military or surveillance are causing concern. In developed western democracies, a social movement is blossoming against such developments. In non-democratic societies like China without such feedback, the technological evolution is progressing without such checks. In the elite technical conferences in AI, China has already achieved top rank in the number of scientific papers published. When compared to the USA or China, India is severely backward in AI. However, fortunately, Indian scientists and students are making rapid progress. The civic society in India is also vigilant about their democratic rights and the possible harm to them from AI. These issues have been clearly outed in the summit on Responsible AI for Social Empowerment (RAISE 2020), organized last week by the Indian government.
In this RAISE 2020 summit, several speakers have pointed out various critical issues that merit deeper investigation in the process of setting up the infrastructure for responsible AI:
- What exactly do we mean by AI?
- Is AI a replacement of man, or is it an assistive tool for man?
- When we examine the work process of AI, to judge whether the produced results are acceptable (for example, according to the 5 criteria of responsibility above), do the Indian citizens need mastery in English? Or can this expertise and technologies be made accessible in Indian languages? To what degree?
- When people submit their agreement for the use of their data in the development of AI, how should the letter of agreement be designed? Indeed, how many people would have the expertise to make an informed judgement on the question of their data? Should we sort the users into various groups, according to their technical knowledge and expertise?
- The current discussion about the ethical behaviour of AI, or the human rights in the use of AI, is being conducted almost entirely within the western philosophical framework. Is it possible to build such a debate on the basis of Indian ethical history and values?
- Owing to the specific socio-economic situation and the enormous human diversity of India, should AI technology have specialized requirements and limits of use within India?
In order to discuss these issues in the right amount of depth, we need to understand the cultural history of “artificial intelligence” in the western world, and how this technology evolved until today. Simultaneously, we need to be aware of the place of Jñāna (knowledge) in Indian philosophical thought, the connotations of the words like Buddhi and Prajñā that are related to “intelligence”, and which opinions exist about the possibility of any of these concepts to be recreated artificially or not. The validity of these opinions in this technological era needs to be also examined.
Intelligence and Artificial Intelligence in Western Philosophical Thought:
The Greek philosopher Aristotle has opined the following in his treatise on politics:
“It is natural and expedient that some men become rulers and the others become ruled. Based on birth and other attributes, some men achieve kingship. The sign of a ruler is the intellect needed to examine any situation rationally. Intelligent men possess this intellect to a greater degree. Hence, it is natural that they rule over women and also those inferior men who have become natural slaves due to their preoccupation with bodily labour. Even worse than all men are non-human animals. Their well-being is ensured only by them being possessed under human occupation.”
Thus, it is clear that Intelligence is being defined by Aristotle as an instrument (means) to achieve and demonstrate power. This perspective has continued for a long time in western philosophical thought. In the 17th century, the French philosopher Rene Descartes ascertained that intellect is limited only to humans, and not available to animals. He argued that only humans possess the consciousness to reason on questions such as “who am I”. Since animals don’t possess this capacity, he argued that they are equivalent to machines. Only humans have the magical property called “Soul” (Aatma). The prominent 18th-century German philosopher Immanuel Kant went further and argued that Morality (Dharma) is limited only to humans who are conscious beings and who possess Rationality (vichāraṇa śakti), and therefore, other beings like animals are “immoral” (dharmaśūnya). The beings which don’t possess rationality don’t have any original value within their own existence, but only a relative value (sāpēksha mūlya) in how they are compared to each other, and how they are being useful to those with original value (humans of a certain kind). It is clear that Kant agreed in principle to the opinion of Aristotle from millennia ago “Some people are rulers, and some people are slaves. Intelligence is the difference between them.” The German philosopher Hegel mentioned that man needs to mould nature using the power of his intellect, and thus shape history. He criticized eastern societies as lacking this capacity to drive history. Marx, who developed the ideas of Hegel further, has opined in a similar manner.
This perspective on intelligence as a tool for power was not limited to philosophical debates. In European history, it has been argued that other races are inferior on account of their supposed lack of intelligence and that thus, they are natural slaves. Such theories of racism enabled the colonization, enslavement, and even the genocide of other races. Severe debates were also conducted whether women possess the right degree of intelligence to merit a “soul”, or whether they deserve democratic representation. Many have argued that women need to be considered as the property of men, owing to their supposed deficient intelligence. Similarly, many discussions happened on the age when intelligence gets expressed in children. Prior to this age, it was argued that children need to abide by the dictates of the adults and that any punishment meted to them was justified. It is only in recent times that the rights of children or animals have become agreeable to the western intellectual community. For many years, they were not recognized on the grounds of supposed deficient intelligence. The subtle underlying cause of the feelings of panic in western civilization about Artificial Intelligence might be due to this perspective about seeing intelligence as a tool for power. This needs to be analyzed.
According to the theories of European philosophers, people started to build intelligent machines in the 18th century, using the new capacities of the industrial revolution. Before that, in the 15th century onwards, clockwork machinery and toys were prepared. Such automata (svachalita vastu) depicted toys with dynamic movements. In 1739, the French inventor Vaucanson prepared a machine known as the “digesting duck”. This duck-machine was able to eat some grains of wheat through its mouth, and then produce pellets of excreta. In reality, the eaten grains were transferred into a compartment within the machine, and imaginary pellets of excreta stored in a different compartment were excreted outwards. Vaucanson was not a mere sculptor but produced several path-breaking inventions. His role in the mechanization of handlooms was very prominent. Such industrial looms powered the European industrial revolution. The automaton of Vaucanson is not a mere toy, but a depiction of a societal perspective, stemming from Descartes. In this view, a real living duck is also nothing but a machine. Seeing the organs of the body through this mechanized view achieved a great revolution in medicine. It was postulated that humans have a spiritual organ called the “soul” beyond the machinery of the body, but other animals do not possess even that. Since these living beings can be seen as machines, their raising can also be industrialized. That is the foundation of the current industrialized animal farming. Such mechanization can also be implemented on forests, rivers, trees, animals and so on, for the unlimited exploitation to the use of man. Due to this opinion, a terrible felling of forests and wildlife happened the world over.
Western scholars also started to see and analyze human societies through this mechanized perspective in humanities and archaeological sciences. In order to analyze how ancient societies lived, scientists started to examine the material artefacts such as ruins, buildings, statues or utensils. These items are called “Artefakto” in Latin. This word means “constructed” (Fakto) through “skill” (Ars). This can be translated as “Naipuṇyīkṛti” in Telugu. It is only from this word “Artefakto” that the sense of “artificial” has been produced. We can translate this is “Kṛtrima”, but then we would already lose the sense of “skill” (Naipuṇya). Behind the phrase “artificial intelligence”, lies the delicate sense that somebody created that object with skill. In the 19th century, it is slowly getting accepted that it is possible to build an artificial structure to hold intelligence, similar to the mechanized engines and automata of the previous centuries. The western scientific community is broadly in agreement that it would be possible to mechanise “soul” – the object that was set aside by Descartes especially for humans. How do we compare this mechanized perspective with Indian scientific tradition and philosophical thought?
Intelligence and Artificial Intelligence in Indian Philosophical Thought:
Knowledge (Jñāna) has an exceptional place in Indian culture. The ideal of human existence is identified as wisdom (Prajñāna), which was marked as the only possible means for liberation. The sacred Advaitic aphorism in the Upanishads “Prajñānaṃ Brahmā” identifies this wisdom with the very basis of the universe. In Buddhist tradition, the achievement of the Buddha-nature was also recognized as possible only through the discrimination by knowledge. By considering the prominence for logical debate (Tarka Vichakshaṇa) in the ancient tradition, and the manner in which debates (Saṃvāda) between intellectuals were conducted with the appropriate decorum, the respect towards knowledge in Indian culture becomes clear. However, the Indian perspective towards intelligence is quite different from the western one. Instead of the focus on authority and power, the emphasis in this culture is towards renunciation and meditative wisdom.
In the Bammanavagga section of the Dhammapādas of Buddhism, Gautama Buddha defines the perfect embodiment of intelligence as a Brāhmaṇa and lists his qualities. The Buddhists consider the Brāhmaṇa nature to be achieved not by birth, but through the practice of meditation (Dhyāna) and discrimination (Vichakshaṇa), by which all the bodily bonds are broken, and through which a person shines in enlightenment. Buddha said that a Brāhmaṇa would break the shackles of desires, be liberated from anger and all hatred, would throw away the burden of ignorance, and become an embodiment of wisdom in his entire being. We can see how contrasting this opinion on intelligence is, when compared to the western perspective of Aristotle, Kant and others.
According to the Sāmkhya system, human consciousness is divided into two primal elements known as Purusha (Perceiving Self) and Prakṛti (Nature). Within these divisions, bodily sensory awareness (Indriya Chaitanya), mind (Manas), intellect (Buddhi), attention (Chitta), memory (Jñāpakaśakti), will-power (Saṃkalpa Bala) and even the awareness of consciousness which says “This is I” (known as Ahaṃkāra in Indian philosophy, and “ego” in English) is all attributed to Prakṛti. Therefore, the “soul” or “intelligence” discussed by western philosophers falls entirely within Prakṛt. Then what is “Purusha”? Set apart from the action-energy and the fruits of action, Purusha is just a witness (Sākshībhūta) of the phenomena that appear in the current moment of awareness (Pratyaksha). The Sāmkhya understanding is that both men and women contain an underlying Prakṛti and Purusha within them, and that not just humans but all beings (Samasta Jīva) possess them. According to Sāmkhya tradition, when one develops attachment (Mamakāra) towards any of the body, mind (Manas), intellect (Buddhi) etc. within Prakṛti, then that would lead to suffering (Śōka). When one discards all of this and identifies only with the Purusha in his Pratyaksha, that state is known as Kaivalya (Aloofness). That is identified as the state of enlightenment for the self (Ātmasiddhi). The Yōga Darśana follows Sāmkhya and defines the state of liberation as Samādhi (Equipoised stability). The Vēdāntins define this as Brahmatva (union with the expanding whole/universe). The Buddhists define it as Nirvāṇa. Although all these practices (Sādhana) have subtle differences, all of them are based on the Kaivalya Sādhana of Sāmkhya, and require a renunciation of sensory attachments and ego. Thus, the western philosophical tradition which declared intelligence as a means to impose authority is opposed to all these systems.
The Sāmkhya system considers each object in nature (Prakṛti) to be composed as a mixture of three Guṇas (qualities): Sattva, Rajas and Tamas. Based on their proportions, any object undergoes transformations and evolves in shape and form. Nothing is permanent. Amongst these 3 Guṇas, while Tamas signifies inertia, laziness and heaviness, Rajas signifies energy, authority and emotional colour. The western philosophical tradition is limited to these two Guṇas. Therefore, they defined intellect as the equivalent of Rajas. However, Indians considered a third Guṇa (Sattva) as aloof from either Rajas and Tamas. Sattva signifies clarity, eternal freshness and the capacity for renunciation (Tyāgaśīlata). It ultimately means the rejection of the gross sensations of the body and focusing the mind towards the subtle inner realm. Thence, Sattva is identified as the signifier of intelligence (Mēdhāśakti). Indians identified Rajas only as the signifier of muscle power (Bāhuśakti).
Indian philosophical thought is diverse, and since immemorial times, had diverse streams: 6 Āstika Darśanas and 3 Nāstika Darśanas. The Āstika Darśanas are divided into three pairs. The Nyāya system is concerned with the definition of objects and the epistemological means (Pramāṇa) for logical debate (Tarka Vichāraṇa). The Vaiśēshika system paired with this is concerned with the analysis of the physical properties of materials in the universe. This pair can roughly be compared with mathematics and physics, although their conjectures would be different from the modern perspective in many subtle ways. The Sāmkhya system which was discussed earlier is concerned with the cosmogeny and the transformations in the universe. The Yōga system which is paired to that is concerned with the liberation of man (or any being/Jīva) from suffering and anger, and elucidates the behaviour needed for the identification of self (Ātmasāyujyata) with immense joy (Mahadānanda). Such logical analysis would help drive the daily habits and practices in Yōga Darśana. Even the Nāstika traditions of Jainism and Buddhism borrow from the Sāmkhya analysis. While the Āstika Darśanas discussed so far are concerned with individual philosophical thought, the remaining Mīmāṃsa and Vēdānta Darśanas are concerned with the philosophical enlightenment of the society and the universe as a whole. The Mīmāṃsa system is concerned with the rituals contextual to space and time in any place and to the pleasing of the creative forces of the universe known as Dēvatas. The Vēdānta system that is paired to this is concerned with the overcoming of all divisions and the path towards the union of a being with the entire universe. This ultimate ideal is known variously in different Vēdānta schools as the non-dual experience (Advaitānubhūti) or proximity to Divinity (Bhagavatsannidhi).
The Chāndōgya Upanishad describes the episode of Sanatkumāra giving the instruction of Bhūmavidya to the sage Nārada. Here, the various layers that constitute intelligence such as the senses, mind, intellect, memory etc. are clearly defined and the distinctions between them are elucidated. The process of naming (Nāma) of different objects is identified as the grossest state of intelligence. From this gross state, various subtle qualities of intelligence are described in the following order: speech (vāk) is superior to name, mind (manaṣ) is superior to speech, will (saṃkalpa) is superior to mind, memory (chitta) is superior to will, contemplation (dhyāna) is superior to memory, understanding (vijñāna) is superior to contemplation, mental strength (bala) is superior to understanding, the water (jala) in which ideas flow is superior to strength, the heat (tējas) that instigates ideas is superior to water, the space (ākāśa) in which all ideas are placed is superior to heat, the recollection (smaraṇa) which can recreate mental objects is superior to space, the hope (āśa) for recreating mental objects is superior to recollection, the life-breath (prāṇa) which maintains the mental time is superior to hope, the unmoving eye of the Self, which is known as Truth (Satya), is superior to life-breath, the blended pair of thought-understanding (matīvijñāna), where thought and understanding are identified as two sides of the same coin, is superior to Truth, the steadiness of faith (Śraddha) is superior to thought-understanding, steadfastness (Nishṭa) is superior to such faith, the meditative action (Kṛti) is superior to steadfastness, the deeper pleasure (Sukha) that is attained by meditation is superior to action, the meditative plane (Bhūma) on which all this is placed is superior to pleasure, the awareness of no-difference (na-anyatva) is superior to the plane, the identification with the most superior Purusha (Aham Purusham) is superior to the awareness of no-difference, and the identification with the all-pervading Ātman is superior to the above identification with the most superior Purusha. We can see that it gets increasingly hard to describe the higher subtle shades of intelligence in English. Based on this instruction in Chāndōgya Upanishad, the process of a man rising to be the embodiment of enlightenment requires turning the focus from the grosser states of intelligence towards the more subtle states and meditating on them.
In this list of the various states of intelligence, Artificial Intelligence (Kṛtrima Mēdhā) corresponds to the grossest state of naming (Nāma), while the user who is interacting with that AI can be considered as activating the second state of intelligence (Vāk). Both of these are good and can be seen as instruments towards the meditative path. However, this path doesn’t stop here and consists of several steps further. Can an AI help towards climbing all these steps? This question might sound laughable. Currently, the various applications are using AI only to distract the human users, not as a means of Tantra (software/meditative framework). But why not? According to the Indian philosophical thought, there is a very lofty ideal for the development of AI. But this is not recognized as such by Indian scientists. When we see AI only through the western perspective, this ideal is not illuminated.
Image: The magical floating island in the sky “Laputa” from the Hayao Miyazaki film “Castle in the sky” depicts the entire island being run by AI machinery.
Continued in Part 2