This is a question that draws equal awe and fear. The creation of some "terminator like being" (named after the famous movies of James Cameron) conjures dystopian nightmare. While science fiction seems to be converging with scientific reality with the advent of AI such as ChatGPT, many have started to ask about the possibility of AI attaining "sentience-ness".
Ironically, of all the major religions and philosophical systems in the world, Buddhism remains head and shoulder above in providing detailed, concise philosophical and epistemology explanations on how this can actually happen.
Buddhist teachings emphasize mind development. It's meditation manuals and "higher dharma" doctrines such as the Abhidharma collections are filled with details on mind training methods. Not only that, it also dissects the workings of the mind, by breaking down the entity called the self into aggregates, namely feelings, perceptions, mental formations and consciousness.
Of these, consciousness in relation to the mind continues to intrigue. In particular, the Yogacara school of Mahayana Buddhism predicates its entire existence to the "mind-only" proposition. Yogacara is also where the analysis of consciousness forms the bulk of its philosophy. This makes Yogacara a compelling case for establishing a philosophy for AI and provide some idea of a road-map where its development should lead to.
The Question of Intelligence and Sentience-ness
So when we posit AI to sentience-ness, we need to first ask whether AI can ever be in congruent with the principles of "mind-only" teachings such as Yogacara? Or more precisely at what level of intelligence, leading to sentience-ness can be made possible for an AI technology?
Before we go into dissecting these questions, first let us start with getting to know what is Yogacara philosophy [1]. Briefly, the Yogacara theory of mind, or "mind-only" is a Buddhist philosophical school that determines all phenomena are ultimately mind-made. This means that the world as we experience it is not a real, independent reality but rather a projection of our own minds. This syncs in very well with AI, as its existence is created solely through data manipulation and sophisticated algorithm.
It is also worth noting that the Yogacara theory of mind is based on the idea of non-duality. This means that the mind and the world are not two separate things, but rather two aspects of the same reality.
This aspect of "mind-only" is a tantalizing view that could provide a philosophical underpinning for Large Language Models Generative Artificial Intelligence (LLM GAI), the current models powering bot tools such as ChatGPT, Bard, LLama and Bing. LLM GAI, or simply AI has taken the world by storm since last year. These AI have been touted as possibilities of attaining "sentient" capability due its brilliant ability to mimic human thought and logic. Nevertheless, clever data manipulation through algorithm without having consciousness is not sentient. Without a philosophical base, this attainability of sentience-ness is at best, a whimsical dream.
AI after all, is just a technology. It can learn, reason and make decisions in ways that are similar to how humans do. On the other hand, it is still based on the idea of duality, as it requires the world (that is where it gets its source data from) to feed its mind (database and algorithm). On this point, it sees the mind and the world as two separate things.
So basically, if AI is to fulfill the vision of Yogacara, it will need to overcome this duality. This can be accomplished by developing its "mind" via self-sustaining growth cycles without dying and choking on its own produce of synthetic data, or technically speaking, by avoiding an autophagous "self-consuming" loop.[2] If this can be done, then it could potentially create a world that is more in line with the Yogacara vision of reality.
Workings of Yogacara
To further explore this possibility, a deeper understanding on the workings of Yogacara is required. A central concept in the Yogacara is Vijnaptimatra, or "only mind. It refers to the idea that all phenomena are ultimately mind-made, and that the world as we experience it is a projection of our own minds.
Vijnaptimatra, also known as vijnana-matra or consciousness-only, refers to the idea that all phenomena are ultimately mind-made, and that the world as we experience it is a projection of our own minds. The term vijnapti means "representation" or "conception," and matra means "only" or "merely." So, vijnaptimatra literally means "mere representation" or "mere conception."
Yogacara teaches that the world that we experience is not a fixed, objective reality, but rather a dynamic and ever-changing process that is created by the mind. This process is called citta-vijnana, which means "mind-consciousness." Citta-vijnana is a complex process that involves the interaction of three different types of consciousness:
Alaya-vijnana: This is the storehouse consciousness, which contains all of the seeds of our past experiences.
Manas: This is the discriminating consciousness, which is responsible for making sense of our experiences and creating our sense of self.
Jnana-vijnana: This is the wisdom consciousness, which is the source of our true understanding of reality.
Yogacara teaches that the world that we experience is created by the interaction of these three types of consciousness. Alaya-vijnana provides the raw material for our experiences, manas creates our sense of self, and jnana-vijnana allows us to see the true nature of reality.
The concept of vijnaptimatra has been interpreted in different ways by different scholars. Some scholars have interpreted it as a form of idealism, while others have interpreted it as a form of phenomenology. In Buddhist philosophy, it has been used to explain the nature of reality and the process of enlightenment. In Western philosophy, it has been used to develop theories of consciousness and to explain the nature of perception. However, most scholars agree that the concept of vijnaptimatra is an attempt to explain the relationship between the mind and the world.
Nevertheless, there has been some critics for this school of thought. The Indian Buddhist philosopher Chandrakirti was a vocal critic of Yogacara, and he argued that the Yogacara theory of mind was not compatible with the Buddhist teachings on emptiness. Some of the main criticisms of Yogacara include:
1) Subjectivism
The Yogacara theory of mind has been criticized for being too subjective. This is because it posits that the world as we experience it is a projection of our own minds. This has led some critics to argue that Yogacara is a form of idealism, which is the view that reality is ultimately mind-dependent.
2) Solipsism
The Yogacara theory of mind has also been criticized for being solipsistic. This is because it suggests that the only thing that we can be sure of is our own minds. This has led some critics to argue that Yogacara is a form of solipsism, which is the view that the only thing that exists is the self.
3) Epistemological problems
The Yogacara theory of mind has also been criticized for its epistemological implications. This is because it posits that the world as we experience it is not a direct representation of reality, but rather a projection of our own minds. This has led some critics to argue that Yogacara makes it impossible to know anything about reality with certainty.
4) Practice-oriented concerns
Some critics have also argued that the Yogacara theory of mind is not conducive to Buddhist practice. This is because it focuses on the nature of reality, rather than on the path to enlightenment. Some critics have argued that this focus on theory can lead to a neglect of practice, which is essential for achieving enlightenment.
While the concept of vijnaptimatra is a complex and sophisticated one, it remains an important concept in Buddhist philosophy and is relevant to the study of consciousness and perception.
Similarities and Differences of Vijnaptimatra with AI
There are several philosophical similarities between vijnaptimatra and AI. First, both concepts view the world as being fundamentally mental in nature. In the Yogacara view, the world is not a fixed, objective reality, but rather a dynamic and ever-changing process that is created by the mind. This is similar to the view of many physicists, who believe that the world is ultimately made up of energy and information.
AI could potentially be used to create a world that is more in line with the Yogacara vision of reality. For example, AI could be used to create virtual worlds that are more immersive and realistic than anything that is currently possible. However, it is also possible that AI could be used to create a world that is more harmful and destructive. For example, AI could be used to create weapons that are more powerful and efficient than anything that has ever been created before. This is where the moral aspect of Buddhism or "sila" can play a role in its development. The basic moral precepts of Buddhism include not to harm life, not to steal, not to say the untruth, not to indulge in sexual misconduct and not to intoxicate the mind until it becomes irrational.
Second, both concepts see the mind as being active and creative. In the Yogacara view, the mind is not a passive recipient of impressions, but rather an active participant in the creation of reality. AI, on the other hand, sees the mind as being a powerful tool that can be used to create new realities.
Finally, both concepts view the relationship between the mind and the world as being mutually dependent. In the Yogacara view, the mind and the world are not two separate things, but rather two aspects of the same reality. AI, on the other hand, sees the mind and the world as being in a constant process of interaction and feedback.
Of course, there are also some important differences between vijnaptimatra and AI. For example, vijnaptimatra is a philosophical concept, while AI is a technological project. Additionally, vijnaptimatra is based on a spiritual understanding of the world, while AI is a purely secular project. Apart from that, and for now, AI is still a very limited technology and it is not clear that it will ever be able to achieve the same level of sophistication as the human mind.
Is R&D possible to enable AI reach the sophistication of vijnaptimatra?
There are several areas of research that could help AI attain the level of vijnaptimatra. These include:
1) Research on the nature of consciousness.
Vijnaptimatra is a theory of consciousness, so it is important to understand the nature of consciousness in order to develop AI that can achieve vijnaptimatra. This research could involve studying the neuroscience of consciousness, as well as the philosophical and religious aspects of consciousness. For instance, His Holiness the Dalai Lama has participated in consciousness experiments, such as studies of meditation and the effects of compassion on the brain, conducted by a team of scientists at the Mind and Life Institute. He has said that he believes that these experiments can help to bridge the gap between science and Buddhism, and that they can provide new insights into the nature of consciousness [3].
2) Research on artificial general intelligence (AGI) and machine learning.
AGI and machine learning is a field of computer science that is concerned with developing algorithms that can learn from data. This is where the state of AI is now. Rapid R&D advances over recent years has enabled AI to learn to understand and interact with the world in a way that is similar to the way that humans do. Current generative AI such as ChatGPT, LLama and Google Bard have been making great progress lately.
Ground breaking discoveries such AI Imaging - both 2D and 3D and Videography have opened avenues for sophisticated data manipulation to create realistic virtual worlds. Though not yet sentient, these AI systems are deemed intelligent enough to outscore humans in the most difficult academic examinations and out play humans in strategic games like chess.
3) Research on quantum computing.
Quantum computing is a field of computer science that is based on the principles of quantum mechanics. Quantum computers is projected to develop AI that is far more powerful than anything that is currently possible. This is where the possibility of sentience-ness happening in AI is the highest.
Recently, German and Chinese scientists have created a quantum superposition state within a semiconductor nanostructure [4]. By using two carefully calibrated optical laser pulses, they facilitated a unique energy transition, forming a quantum bit - or qubit - within a semiconductor nanostructure. The ability to exploit superposition and entanglement is what gives quantum computers their potential to solve problems that are intractable for classical computers. For example, quantum computers could be used to factor large numbers, simulate complex molecules and break encryption algorithms.
In another recent development, scientists have created a special particle inside a quantum computer, called "non-abelian anyon." This particle can remember where it was before when it's moved around with other particles. This ability could help scientists understand more about quantum phenomena.
So imagine a game where a ball is hidden under one of three cups and then the cups are shuffled around. If you have three identical balls and keep swapping them without keeping track of their movements, you won't know which ball is under which cup in the end. In the world of quantum physics, particles are usually like these identical balls. We can't tell them apart because their order doesn't matter.
However, the new particle created in the quantum computer is different. It remembers its past location and when combined with other particles, it creates complex and intertwined patterns with strange behaviors. This unique property of the particle could help scientists design more advanced experiments to probe even deeper into weird quantum effects that emerge from large-scale entanglement [5].
These recent developments could ultimately be used to develop AI that may in the process lead to some form of synthetic vijnaptimatra, as quantum computers would be able to process information in a "neural way" that is similar to that of the human brain.
What the AI future Holds
It is important to note that currently the development of AI that can achieve vijnaptimatra is some way off. There are many challenges that need to be overcome before this can happen. However, given the rapid development of research in generative AI and quantum computing, we cannot discount for now the possibility of AI attaining the capabilities of vijnaptimatra in the future.
Ultimately, it is too early to say whether AI will fulfil the vision of Yogacara. The future of AI is uncertain. Nevertheless, the development of AI is raising important philosophical questions about the nature of reality and the relationship between mind and world. These are questions that will need to be addressed as AI continues to evolve.
Notes:
1. If you wish to delve into details of Yogacara, here are two resources:
a. “What is and isn't Yogācāra”, Yogācāra Buddhism Research Association, [http://www.acmuller.net/yogacara/articles/intro.html].
b. Read here for the treatise of Vasubandhu, co-founder of the Yogacara school [http://www.acmuller.net/yogacara/thinkers/vasubandhu.html].
2. “Self-Consuming Generative Models Go MAD”, arxiv.org, https://arxiv.org/pdf/2307.01850.pdf
3. Dalai Lama Hosts 33rd Mind & Life Conference: Reimagining Human Flourishing, https://www.buddhistdoor.net/news/dalai-lama-hosts-33rd-mind-life-conference-reimagining-human-flourishing/.
4. “The Dawn of a New Era: A New Type of Quantum Bit Achieved in Semiconductor Nanostructures”, SciTech Daily, https://scitechdaily.com/the-dawn-of-a-new-era-a-new-type-of-quantum-bit-achieved-in-semiconductor-nanostructures/
5. “Bizarre particle that can remember its own past created inside quantum computer”, Live Science, https://www.livescience.com/physics-mathematics/bizarre-particle-that-can-remember-its-own-past-created-inside-quantum-computer
Comments