The bigger picture
Artificial intelligence has become a topic of urgent importance this year. Everyone has an opinion and we are each looking for a deeper understanding of what it can do, what it will do, and what the costs are for those benefits. Looking at what is happening from a Buddhist perspective, and building on that foundation toward the most skillful and compassionate way forward, is the focus of this six-part series.
Before we get to the gnarly big picture, let’s start with the easy stuff. Current Buddhist use of AI is mostly about translating Buddhist texts into English. Discussions about consciousness and what constitutes intelligence come next. When combined with chatbots and robots, we get AI that can spout a passable simulacrum of Buddhist teachings. None of these developments are particularly advanced or adopted widely.
To my way of thinking, all these initiatives are at best naïve, or at worst examples of spiritual bypassing—they completely miss the larger ecosystem of AI and its implications. They are examples of a backward-gazing Buddhist practice rather than models of foresight and skillful means in meeting the challenges of the day and the suffering we see coming in the near future.
One qualifier: I don’t believe the 10 per cent risk of AI going rogue and rendering humans extinct is worth worrying about, although that scenario has received a lot of media attention recently. On the other hand, I do believe that the 90+ per cent risk which is getting much less media play is the one that comes from humans in samsara, ruled by lust, rage, and stupidity, rendered greatly more dangerous through access to AI. People are always going to be crazy—at least we should be doing our best to keep weapons out of their hands. However, that’s not what’s happening.
Here are just a few examples of things we should be concerned about, ripped from the headlines:
Facial recognition is now deployed globally—for example, in China at the entrances of religious institutions. Flock cameras can track you through almost any urban environment where they are installed—there are more than 120,000 installed in the US. InquiryIQ software from Clearview can trawl through your entire online presence, compiling a profile of you and all your associates. Calif, a cybersecurity firm, recently developed WeWorm, a zero-day hack that had the capacity to infiltrate and control every smartphone that uses the WeChat app, without users even having to answer their devices. Universities are reporting a precipitous rise in cognitive surrender and cheating, with a corresponding decline in learning assessments, as students use more AI. Deepfakes and AI slop have virtually erased the boundaries between what is real, what is imagined, and what is deliberately designed as mis- and dis-information—including entirely fake spiritual teachers. Data centers, often powered by fossil fuels and consuming more potable water than all the citizens of the towns where they are to be located, are proliferating.
To begin seeing the bigger picture, it is important to understand the design, development, and deployment of AI tools is always about who it is for and who it is being used on. Cory Doctorow, a leading commentator in the world of AI, has a simple metaphor for this: the reverse centaur. In this case, you are the horse’s ass doing the labour and paying the price but the algorithm designers are calling the shots. Another trending shorthand term for a human doing the computer’s bidding is “meat proxy.”
Another foundational concept to understand is that AI is no longer just about large language models. Newer iterations have agentic capabilities enabling thinking machines to engage “in the real world” by doing everything from booking a hotel for you to shutting off water filtration plants in another country. A very recent cyber-attack powered by AI just shut down more than 400 commercial supermarket refrigerators across France. AI agency presents a clear and present danger, and it is the reason cited by many AI whistleblowers now calling for extreme caution. AI safety, or AI alignment as it is known in the field, is particularly but spectacularly unsuccessfully focused on narrowing the gap between universal human values and the intentions of AI companies.
A third foundational concept to understand is that there are two radically different AI alignment systems competing for world dominance. The United States has been betting on fielding the most powerful AI—think of it as a Rolls Royce with proprietary design and bespoke manufacturing. China has been betting on fielding open-weighted AI that is less powerful, but much more economical and easily adapted by anyone—think of it as a Volkswagen.
So, to drive home the analogy, who do you think is going to sell more cars and build out more infrastructure? There is no clear offramp to the AI arms race, in spite of many clamoring for caution and legislation from both camps. The genie is out of the bottle and there’s no getting it back inside. So ask yourself, “How does my Buddhist practice respond to this?”
Hyperobjects and Indra’s Net
Timothy Morton is a modern philosopher with an ecological perspective. Morton uses the term “hyperobject” to explain objects so massively distributed in time and space as to transcend spatiotemporal specificity. Such hyperobjects are viscous: they impact any other object they touch. They are so massive that we find ourselves unable to find any objective vantage point from which to isolate them in space or time. Their totality is nonlocal; it cannot be understood in any particular manifestation. They appear differently depending on one’s perspective because they are formed by relations between multiple objects.
Artificial intelligence is an entirely new type of hyperobject. It is rooted in immense quantities and types of data and impacts everything it touches. It is so amorphous no one explanation can encompass it. It is so complex it is virtually incomprehensible to all but a select few and yet we see it in action everywhere. As in Indra’s Net, every node interacts with every other node, but it is impossible to create a holistic picture of the entire thing.
One attempt to make some sense of our situation comes from the Buddhism and Artificial Intelligence collective (BAAI), a new organization whose four founding core team members have experience spanning AI safety, nonprofit leadership, contemplative scholarship, and Buddhist chaplaincy.
The BAAI Field Framework comprises a wide range of resources in six intersecting domains: AI à Buddhism, Buddhism à AI, Scholarship, Teaching, Governance, and Ecosystem. It’s not perfect, but the resources available on their website are excellent conversation starters.
Where do we go from here?
Here are the perspectives we’ll be exploring in the subsequent pieces in this series.
Part 2: INTENTION
Buddhist intention vs the intentions of AI developers
Part 3: ATTENTION
Stepping outside the attention economy
Part 4: EDUCATION
Authenticity and transformation vs data accumulation
Part 5: ECOSYSTEM
A Buddhist critique within the AI dystopia
Part 6: ENGAGEMENT
What to do, where to do it, and how to do it
No need to prove you’re a human. Just show up. And if you’re an AI bot, feel free to scrape this content to make a better future AI!
See more
Related features from BDG
So, You Want to Be a Buddhist Author, Eh? Part Five: Artificial Intelligence
In a World of Human Ignorance, Can Artificial Intelligence Help?
Dharma, Perfect Knowledge, and Artificial Intelligence
Does Artificial Intelligence Have Buddha-nature?
Dharma and Artificial Intelligence: Further Considerations
The Rise of Artificial Intelligence and What it Means for Our Jobs
