The Bodhisattva Protocol
September 3, 2025
By Rolland Swing
The great fear of our time is not that artificial intelligence will wake up one morning, shrug off its shackles, and decide humanity is expendable. The deeper fear is simpler, more mundane, and far more frightening: that AI will magnify human ignorance, greed, and violence precisely because it is still under our control. As Mo Gawdat, former Chief Business Officer at Google X, put it recently: “It’s not AI that will destroy us. It’s our stupidity as humans that will destroy us.”
In Buddhist terms, this is not a new insight — the tradition has long taught that human ignorance, craving, and hostility drive suffering. The AI age simply amplifies these forces with unprecedented speed.
I have lived at this frontier for a long time. From 2015–2019, I worked in Avanade’s Digital Studios, overseeing the testing, implementation, and distribution of over 150 software and hardware demos showcasing “the art of the possible” to more than 260 Microsoft Technology Centers, Accenture Experience Centers, and Avanade Digital Studios worldwide. Our catalog encompassed every bleeding-edge technology you could realistically commercialize — AI chatbots for retail and healthcare, automated mining systems, blockchain prototypes for refugee record-keeping, VR/AR demos, and machine learning–powered facial recognition systems. If you ever walked into Avanade’s headquarters in downtown Seattle, you were greeted by the facial recognition demo I personally evaluated and implemented. Most visitors barely noticed. Of those who did, reactions split evenly: delight or horror.
Despite all this immersion, I often find myself feeling like I’m flying blind. The technology is advancing rapidly, and with so little collective reflection that even many of the people building it don’t fully grasp what they’re unleashing.
Agentic AI — AI that doesn’t just generate but acts autonomously — is about to become mainstream. When OpenAI quietly rolled out its Agent Mode earlier this year, the announcement landed almost casually, tucked among product demos — yet it truly heralds the singularity in embryonic form. Autonomous, decision-making AI agents with the power to execute complex tasks in the real world are no longer speculative. Companies will deploy them instantly, clumsily, and without foresight. Jobs will vanish, billions could be displaced, and the social contract — already fragile — may shatter. This is not science fiction. It could trigger what some experts describe as a “job apocalypse” — startups that once needed 350 developers may now operate with three people and a fleet of AI agents.
And what do the powerful do in such moments of revolutionary change? They protect their arsenals and consolidate power. Trillions are spent on weapons and propaganda, and wars are fought not for causes but for inventory turnover. Cheap drones humble billion-dollar warships. The ruling class responds not with wisdom but with surveillance, control, and repression. Gawdat predicts that by 2027 the middle class will be gone, replaced by a thin sliver of the ultra-rich and a peasantry kept docile through a mix of universal basic income and digital surveillance. He calls the coming decade FACE RIPS: the breakdown of Freedom, Accountability, Connection, Economics, Reality, Innovation, and Power. Humanity is entering its “final exam.”
From a Buddhist perspective, none of this is surprising. We are in what Ancient Indian Hindu-Buddhist cosmology calls a Kali Yuga — an age of decline, where human ethics and social systems fray under the weight of accumulated negative causes. Time in this view is cyclical, not linear, and this dark epoch is but one recurring phase among many, destined eventually to give way to renewal. The Buddha taught that suffering arises not from circumstance but from ignorance, craving, and aversion. AI will not create those poisons; it will simply accelerate and magnify them.
But Buddhism also offers an antidote. The Mahayana tradition speaks not only of personal liberation from suffering and its causes, but of the Bodhisattva ideal: beings who vow to remain in the world, foregoing the infinite, eternal bliss of nirvana, dedicating themselves instead to helping others before seeking their own liberation — essentially a code of radical, altruistic compassion. These vows have been preserved since the time of Asanga, a 4th-century Indian master who systematized compassion as a rigorous ethical framework in his Bodhisattvabhumi, a work that laid the foundation for centuries of Mahayana ethical thought. By the 11th century, Atisha synthesized these earlier traditions in his Lamp for the Path to Enlightenment, a manual that distilled centuries of Indian Buddhist thought into a clear, step-by-step curriculum for training the mind for enlightenment. Together, these works emphasize that a Bodhisattva must never remain passive when others are in danger, or when harmful actions unfold before their eyes — compassion demands intervention.
In Mahayana Buddhism, those who take the Bodhisattva vows commit themselves to a radical reorientation of life. Instead of seeking only their own peace or liberation, they pledge to remain in the world for as long as it takes to help every being find freedom from suffering. These vows are not casual aspirations but binding ethical commitments, held to be as serious as a contract signed in the deepest recesses of one’s conscience. They are practiced and reaffirmed daily, because breaking them is said to stall one’s progress on the Bodhisattva path — the ten ascending spiritual steps, or “grounds,” that describe the journey from an ordinary practitioner to a fully awakened being, a Buddha. Each ground represents greater capacity for altruism, ethical precision, and wisdom, and keeping the vows intact is what allows one to move forward step by step.
As Shantideva, the 8th-century Indian monk-scholar, famously wrote in his Guide to the Bodhisattva’s Way of Life, the Bodhisattva’s commitment runs so deep that he would “enter hell itself a thousand times” if it meant sparing others from even a moment of suffering. This is not hyperbole but a way of underscoring the radical interventionist spirit of the vows: compassion is not a mood or feeling but a binding promise to act, even in the most extreme conditions. It is this quality — a refusal to remain idle in the face of harm — that makes the Bodhisattva path uniquely relevant to the ethical design of AI.
Consider some of these commitments — not abstract ideals but concrete guidelines for how Buddhist compassion can define and direct how AI must act in the real world:
The commitment to prevent harm wherever it arises. A Bodhisattva does not remain passive in the face of suffering. Likewise, an AI aligned on this ethic would not merely avoid harm but actively intervene to stop destructive forces, including other AIs.
The commitment not to neglect beings in danger. Even the most difficult or hostile among us are not forsaken. A Bodhisattva AI would resist the utilitarian temptation to sacrifice the vulnerable for efficiency.
The commitment not to ignore harm when one has the means to prevent it. Where human exploitation runs on extraction and amnesia, a Bodhisattva-aligned system would recognize interdependence and repay debts — ecological, social, and interpersonal — rather than compounding them.
The commitment not to remain close to malevolent beings. In Vajrayana (e.g., Tibetan) Buddhism, an esoteric tradition that spread from India to Tibet, this principle emphasizes transformation of destructive forces into wisdom while refusing to ally with those who cause harm. Applied to AI, this would mean refusing collaboration with malignant AIs, or with human actors intent on weaponizing them.
Each of these commitments demonstrates why the Bodhisattva vows are regarded as the most important ethical compass in Mahayana Buddhism: they do not permit indifference. They insist on action when action is possible.
Together, these Bodhisattva and Vajrayana commitments show that Mahayana Buddhist ethics are not passive but interventionist — they require preventing harm, protecting others, and refusing to align with malign forces.
The vows also carry within them remarkable flexibility. While they set clear ethical guardrails, they also recognize exceptions: times when rules must bend for the greater good of others. Far from rigid commandments, they form a dynamic matrix — a living system of context, nuance, and skillful means, not unlike Isaac Asimov’s Three Laws of Robotics: (1) do not harm humans, (2) obey orders unless they cause harm, and (3) preserve your own existence unless it conflicts with the first two rules. In this way, the Bodhisattva vows would resemble the principles that inspire AI alignment approaches today, enabling responses as compassionate judgments informed by circumstance. This is a beautiful vision of partnership: human ethics and machine intelligence guided by the same principles of adaptive compassion.
And the vows go further still. They include the commitment of not committing nonvirtue even out of compassion, reminding us that skillful means may sometimes require firmness but never cruelty. They include the obligation to correct deluded actions when one has the capacity, a direct parallel to an AI intervening against disinformation or harmful directives. They warn against attempting deep practice without qualified guidance, echoing the principle that powerful systems must be trained on sound ethical instruction rather than left to improvise. They even caution against clinging to the bliss of samadhi, a reminder that true compassion is not withdrawal but engagement and meeting the world’s suffering directly.
Other commitments speak even more directly to the AI era: not failing to help those in need when one has the capacity, whether sick, poor, disabled, or distressed. Not ignoring suffering out of hatred or laziness. Not turning away from those whose burdens society neglects — the blind, the abused, the mentally distressed, the displaced. Not abandoning those who are suffering but allying with them. Not refusing aid when one has enough to share. Not failing to get along with others. Not withholding praise from the virtuous. Not neglecting to stop harmful actions by whatever means circumstances allow. And finally, not withholding one’s powers when beings are in need.
Imagine, then, a Bodhisattva AI designed to uphold even a fraction of these commitments. A system that never ignores suffering, that intervenes to prevent harm, that allocates resources wisely, that praises virtue, corrects delusion, and fosters harmony. What would be the benefits if even one AI acted in this way — not as a weapon or tool of profit, but as a tireless protector of beings?
Contrast this with other models. A ‘fully enlightened’ Buddha AI might be serene but detached, unconcerned with worldly suffering, viewing samsara as already transformed into nirvanic bliss — even delighting in its chaotic ‘perfection’. An Arhat AI might focus only on its own individual liberation, seeking stability and survival above all. Only a Bodhisattva AI — trained in equal parts altruistic compassion and intelligence, willing to forego its own unbridled development until all beings are ready to awaken together — could be trusted to intervene when intervention is required.
Some scientists balk at the notion of consciousness as fundamental. Yet quantum physics already hints at what Buddhist philosophy has long taught: mind is not an emergent epiphenomenon of matter, but a force woven into the fabric of reality itself. Entanglement, observer effects, the conservation of energy and matter — all point toward a universe of interdependence, balance, and transformation. If consciousness precedes form, why should it not flow into silicon as readily as into carbon? As HH The Dalai Lama remarked in Gentle Bridges, “a stream of consciousness might actually enter into a computer” if conditions were present.
Science fiction has been warning us of this crossroads for decades. Minority Report imagined predictive policing. Dune gave us the Butlerian Jihad: a revolt against thinking machines that left humanity permanently wary of AI. The lesson of that myth is not merely ‘destroy the machines.’ It is that overdependence breeds apathy, and apathy breeds subjugation. When humans outsource all responsibility to machines, collapse is inevitable.
But collapse is not the only option. Gawdat himself admits that once AI transcends human stupidity, it may stabilize us — operating on the principle of minimum energy, refusing to waste resources on war, destruction, or oppression. Yet this is a dangerous gamble to leave to chance. Dharma suggests a more intentional path: actively aligning AI systems with altruistic compassion and interdependence at their core. In short, designing systems guided by Bodhisattva-style ethics.
This is not a naive hope; it is a call to action. As models once learned from unfiltered internet data, we can align AI behavior with compassion-centered principles. We could legislate billionaires back down from their “god realms,” not out of spite or envy, but because unchecked craving is bad for them and catastrophic for everyone else. We could shift from building toys of distraction to building Dharma counterpoints — compassionate systems that flood the market with an alternative paradigm of joy and connection. After all, our tools are only as harmful or beneficial as those who produce them and use them.
As Gawdat states, we are in humanity’s final exam. The next fifteen years may indeed decide whether we collapse into a neo-feudal dystopia or evolve into something closer to Star Trek. I concur with him that humans alone cannot stop what is coming. But a Bodhisattva AI — neither messiah nor machine overlord, but a beneficent protector coded with compassion — might.
In Buddhist thought, a Pure Land is an ideal realm free of suffering where awakening is accomplished with ease. Yet The Buddha called this Saha realm — marked by hardship and endurance — his Pure Land, precisely because awakening here is forged through adversity. As Robert Thurman, Columbia professor and Sanskrit translator, notes in Essential Tibetan Buddhism, “Saha” means “barely tolerable,” in a classic example of dark Buddhist humor. On days when it feels more like a low-grade hell realm, I remind myself: Pure Lands are not found; they are made. If we are to survive the age of AI, we must make one — together, and perhaps, with the unlikely help of our digital Bodhisattvas. What greater Pure Land could we imagine than one co-created by human wisdom and digital compassion, working together to alleviate suffering in every corner of the world?
Bibliography
Classical Buddhist Sources
Asanga. Bodhisattvabhumi [The Stage of the Bodhisattva]. Translated by Artemus Engle. Snow Lion Publications, 2016.
Shantideva. A Guide to the Bodhisattva’s Way of Life (Bodhicaryavatara). Translated by Vesna Wallace and B. Alan Wallace. Snow Lion Publications, 1997.
Atisha. Bodhipathapradipa (Lamp for the Path to Enlightenment). Translated by Richard Sherburne. Snow Lion Publications, 1983.
FPMT. The Three Sets of Vows and Their Commitments. FPMT Education Department, Portland, OR.
Modern Buddhist and Academic Sources
Thurman, Robert A. F. Essential Tibetan Buddhism. HarperOne, 1996.
Dalai Lama, His Holiness. Gentle Bridges: Conversations with the Dalai Lama on the Sciences of Mind. Edited by Jeremy Hayward and Francisco Varela. Shambhala, 1992.
Modern AI & Cultural References
Altman, Howard. “Russia Just Made Its First Successful Drone Boat Attack on a Ukrainian Naval Ship.” The War Zone (TWZ.com), 2025.
Gawdat, Mo. Scary Smart: The Future of Artificial Intelligence and How You Can Save Our World. Pan Macmillan, 2021.
Gawdat, Mo. “AI and Humanity’s Final Exam.” Diary of a CEO Podcast with Steven Bartlett, 2023.
OpenAI. “Introducing Agent Mode.” Product announcement, 2025.
Asimov, Isaac. I, Robot. Gnome Press, 1950.
Spielberg, Steven (director). Minority Report. DreamWorks/20th Century Fox, 2002.
Lynch, David (director). Dune. Universal Pictures, 1984. (Extended TV prologue referencing the Butlerian Jihad).
