Prolegomena to any Future Artificial Intelligence: The Teleological and Practical Foundations of the World Integration Loop
by Ryan Boudinot and Julia Mossbridge, Phd
Abstract:
Just as an individual user can submit a prompt to a large language model like ChatGPT, Midjourney, or Gemini, civilization as a whole is collectively offering an unexamined meta-prompt that, for better or worse, represents the warp and woof of humanity’s imagination, intellect, neuroses, and aspirations. Here we consider AI as a superdisruptor–an exponentially accelerated iteration of the much slower forces of disruption to which we have become accustomed in the past century of technological evolution. What is humankind ultimately asking this superdisruptor to do? What kind of world do we want it to shape, and what kind of world is it shaping now? What must our meta-prompt be, even if we are not conscious of its subtext? The answers start with recognizing that even a potential multiverse of optional futures still only leaves us with a single future to eventually experience. In this “time-traveling” project, we discover that a crucial purpose of AI may be to support time-perspective flipping, a practice that allows us to understand meta-prompts driving the timing and nature of superdisruptors themselves. Time-perspective flipping creates a liaison between the physical and representational, binding the biosphere to the digital noosphere through a platform of feedback mechanisms we call the World Integration Loop (WIL). The WIL, first articulated in the IRCAI 2021 Global Top 100 list, operates with a “bird’s eye view” of a temporal map of events, tying our unfolding experience of the real world to digital twins of those events. We provide a schema of the WIL that includes an integration of technologies such as cloud computing, data lakes, digital twins, remote sensing, drones, geospatial data technologies (GIS, space-time cubes, LiDAR, IoT), quantum and fungal/bio computing, blockchain, and game engines. We propose that the underlying meta-prompts emerge naturally from, and also cause the initiation of, the WIL itself. Taken together, what arises from this exploration is a non-dualist reinterpretation of the relationship between nature and technology that seeks to dismantle artificiality as a concept and incorporate what we currently call AI into a conscious, living planet.
Prolegomena to any Future Artificial Intelligence: The Teleological Foundations of the World Integration Loop
AIs as superdisrupting agents
The scientific origin story of humanity tells us that sixty-six million years ago, an asteroid about six miles in diameter slammed into the Yucatan Peninsula, wiping out three quarters of all plant and animal species on earth. In the wake of what has come to be known as the Cretaceous-Paleogene extinction event, one species of mammal, Homo Sapiens, extended itself into a plethora of technologies, including: a global electronic system capable of near instantaneous one-to-one and one-to-many communications; satellite telescopes that scan the heavens for exoplanets, incoming threats, and light from stars as old as 13.5 billion years; processes to manipulate the fundamental building blocks of life; colliders that slam together individual particles of matter; and artificial intelligence technologies that extend creative and analytical properties of the human mind that until now have been considered the sole province of neurons, gray matter, and daydreams.
Regarded through a Western anthropomorphic lens, the story of how humans gave rise to the Internet, generative AI, the James Webb Space Telescope, the CERN collider, CRISPR, and the Swiss Army knife–to name but a few innovations–is one of human triumph over the elements and dominion over the earth. But if one were to tell the same story from the earth’s perspective, what emerges would likely become a tale of a rapidly evolving planetary immune system rising up from eminently disposable civilizations, engineered to mitigate against another ecological catastrophe of the magnitude that relegated the empire of dinosaurs to the fossil record some sixty-five thousand centuries ago.
Here we attempt to contextualize recent developments in the fields of artificial intelligence, machine learning, cloud computing, and geospatial data within the 3.7 billion-year scientific story of life on earth. We’re disinterested in entertaining AI within the narrow window of late-stage neoliberal capitalism, as an opportunity to enrich a select minority of men, or as the foundation for business plans. We propose considering the emergence of AI as analogous to the eruption of a climate-altering super-volcano–a superdisruptor–an agent with the capacity to speedily overturn “controlled” attempts to alter the course of humanity. Similar to how the Tralfamadorians from Kurt Vonnegut’s Sirens of Titan altered humanity’s trajectory over thousands of years to create space travel–just so that they could deliver a screw to the planet Titan to fix one of their robots–locating artificial intelligence upon a panoramic canvas that includes the entire history of life on earth challenges us to imagine an ultimate purpose for its existence in terms that invoke the persistence of life beyond humans and even the earth itself.
We cannot yet know whether AI emerged for some teleological purpose, as a necessary historical event, or both. But in the presence of any superdisruptor that we allow to exist with us on earth, the responsible thing to do is to ask why we are allowing it to exist. The answer is, of course, that we want it to change the world–which means, to change us. That’s why we allow any new innovation to arise and be promulgated. Then the question becomes: how do we seem to want to change? This is not a conscious act, but a deeply unconscious one, so the question requires some digging. We need evidence rather than opinions, if possible. The other question that we are responsible for asking in the presence of any superdisruptor is how should we want to change, if our goal is to achieve the most beneficial outcome for humanity and the planet. After addressing this question, there is the work of making plans to move the motivations governing our AI-human interactions from where they are now to where they must become, assuming we want to experience the most beneficial outcome for humanity and the planet.
How the Meta-prompt Manages Superdisruptors, How Superdisruptors Create Meta-prompts
We are thinking about AI in a beginning-to-end or “all at once” temporal landscape of events in time, but we are also talking about the unfolding process of shaping the future with respect to AI. There is no paradox here; we integrate these two seemingly opposing views of AI and its location and influence on the temporal landscapes by acknowledging that any influence AI will have on the Gaian future will necessarily be viewed by human conscious experiences as an unfolding influence, moving forward in time. Meanwhile, transcending individual conscious experience gives us the “bird’s-eye view”–equally real, just a different perspective–on how AI is situated in history.
Switching between these two views allows us to realize that we do not have to worry about whether a “multiverse” of potential futures exists. We can choose to see the bird’s-eye view map of the past, present, and future of the universe we will end up experiencing regardless of whether there are other universes around. We say this because we are working with the assumption that we want to experience the most beneficial outcome for humanity and the planet–with “outcome” defined as summed across the entire timeline of the universe we end up experiencing, from beginning to end. Given this assumption, it becomes useful to flip our perspective back and forth between the forward-directional-on-the-ground experience of this most-beneficial outcome and the bird’s-eye view map of the predicted (and in this view, actual) path. The more we conduct that flipping practice, the more we can see the relationships between what we think and experience in the conscious mental realm (which unfolds directionally forward) and the physical landmarks on the bird’s-eye-view map of time (Figure 1).
We consider unconscious intentions to be the essential bridge between the conscious mental experience of unfolding events and the physical landmarks on the bird’s-eye view map. The unconscious mind can be considered to live in both perspectives, or different perspectives at different times (Mossbridge 2015). The intentions arising from the unconscious can trickle “up” to consciousness and appear in the context of AI as temporally unfolding “meta-prompts,” or seemingly external causal instructions, for AI and the world. They can also trickle “down” to physical reality and appear, in the bird’s-eye-view-of-time, as meta-prompts that have always influenced the ever-present relationships between human and machine.
Here’s an example. Let’s imagine the on-the-ground conscious experience of the driver of a truck on a road who is approaching a hairpin turn. For the future outcome to be positive, the driver will presently need to experience that the hairpin turn is approaching and that they change their behavior more drastically (make a sharp turn and slow down) than they have in the recent past. Here, the truck driver is the positive disruptor–they are used to doing the good work of shipping many goods all at once. The hairpin turn is the superdisruptor–it’s a paradoxical challenge. It was built by humans as a meta-prompt for the land, to allow a future disruptor to navigate around some physical constraints and also force major adjustments to be made in the disruptor’s approach to driving (Figure 1, left). Let’s look at the same event viewed from a bird’s-eye-view of a landscape of time, showing the disruptor in the middle of its current path, headed for a superdisruptor. In this view, we can see smaller vehicles (lesser disruptors) who must change their behavior as well, but less dramatically than the disruptor, who takes up more time and space. We can also see that humans have built a series of superdisruptors that cause direction to be changed radically in the moment, but we also see from this bird’s-eye-view map of the road and the land that the creation of the road itself is a present-time meta-prompt that moves the entire system of disruptors eventually in one direction or another. This meta-prompt could be a circle, but in fact it is linear, and has some logic embedded in the system itself, including the knowledge of the disruptors, about what kind of disruptors will go toward one direction or the other. Now moving back to the on-the-ground picture (Figure 1, left), we cannot help but embed knowledge of the meta-prompt into the disruptor who in this moment only sees a single superdisruptor ahead. We propose that time-perspective flipping supports self-transcendence like the “overview effect” does. The overview effect was reported by astronauts who were of course experiencing unfolding time as they observed the earth holistically from above.
Now–let’s return to the idea of AI as a superdisruptor. If it is, then we have unconsciously built AI with the purpose of guiding disruptors through external circumstances that require us to change in order to achieve a positive outcome. There are cognitive, emotional, and behavioral outcomes of this perspective-flipping process as applied to AI. Cognitively, the perspective-flipping embeds the meta-prompts into our understanding of unfolding developments in the AI space. Emotionally, it allows us to understand our own “programming” as disruptors to interact responsively relative to meta-prompts because we ourselves created the goal of the system even if we do not, in unfolding time (Figure 1, left) know what it is yet. Behaviorally, perspective-flipping allows us to better navigate the future by calculating beneficial responses to superdisruptors. It is our sense that these changes have been, are, and will be induced by the World Integration Loop, exactly because it induces time-perspective flipping as a matter of course. It is also our sense that this time-perspective flipping may be a core purpose of AI; that is, the meta-prompt we are moving towards is one which the temporal perspective of humanity becomes more inclusive, with fewer boundaries in time.
A Bidirectional Bridge between Worlds: The World Integration Loop
To live as a human in an industrialized society in the early twenty-first century is to conduct three interrelated but separate lives simultaneously–an interior mental life, a physical life beholden to our bodies, and a digital life dictated by the extensions of our bodies and our minds within our electronic devices and platforms. Our generation’s challenge is to align our own digital behavior, expressed through our devices and platforms, to positive outcomes in physical reality. For an expanding cohort in the industrialized world, this means earning money through knowledge work, converting behaviors expressed almost exclusively online into food, shelter, and medicine.
The emergence of generative AI into popular culture in 2023 was striking not simply for the plethora of images and texts that suddenly dominated our social feeds. The most curious aspect was the emergence itself–this was the first time we had observed computers exhibiting behaviors that they had not been explicitly asked to perform. Suddenly it seemed that machines were behaving more akin to ant colonies or flocks of birds than according to the linear and predictable dictates of command lines and circuit boards.
This suggests that we can’t simply ask AI to achieve predetermined functions, as has been our custom since the dawn of the ENIAC, the era of expert systems. Instead, we’re entering an era in which we recognize that we have already established conditions that are conducive for behaviors to emerge, with the expectation that the results may surprise or confound us. Moving from prescriptive, instructive modalities of computer programming to this more open-ended, speculative engagement with our machines situates us more in the role of conjurers than engineers.
So what is it we’d like to conjure? To answer this question requires a potentially painful process of soul searching as we seek to understand the cumulative meta-prompt, influenced by every submission we make to our bots. It is time to articulate the world we want to live in. Just as it is no longer possible to engage with computers as boxes into which we pour instructions, but instead as mechanisms that can respond unpredictably and surprisingly to our prompts, engineering the superdisruptor that can inspire the emergence of humanity’s desired world means approaching the project less as a series of commands and more as the establishment of certain conditions and tendencies.
In 2020, the World Integration Loop (WIL) was concerived as a series of protocols designed to establish positive feedback loops between behaviors like gaming and the purchase of digital goods and measurable environmental action including reforestation, conservation, invasive species management, and coral reef restoration. The WIL is a superdisruptor through which behaviors in digital platforms trigger positive environmental outcomes in the physical world, and these positive outcomes generate data that is in turn ingested into the digital platforms. As we conceive of a bridge between our digital, mental, and physical selves, we can conceive of a bridge on a far larger scale, that of a civilization’s collective digital behavior as a platform for monitoring and enhancing the environmental health of the planet. The WIL is presented as an overview, allowing zooming into the world’s condition in a holistic, forward-and-backwards-in-time manner — a player can project themselves forward in time and see what will happen if nothing changes, and move backwards in time to see what occurred to make things as they are. As we imagine flipping between the physical and digital worlds, we habitually flip between a “bird’s eye view” (digital twin world) and a strictly unfolding (physical world) sense of time. The gift of the WIL is that both directions of perspective flipping are supported by bidirectional causal influences between the physical and digital worlds.
One proposal for the WIL was included in the IRCAI Global Top 100 list of 2021, and served as the theoretical basis for a hackathon-winning project at Google Cloud managed services partner SADA that year. The hackathon project took the form of a Super Mario-style game that integrated reforestation data with a game world. As trees were planted in the real world, digital representations of those trees would appear in the game. Revenue generated by the game would, in turn, be directed to reforestation efforts in the real world that included planting and geo-tagging trees.
The WIL includes games, digital twins, and digital goods and services that create revenue streams to fund climate action activities in the field. In exchange, funded partner organizations provide the data used to build and influence digital platforms. Achievements in restoration, conservation, and stewardship initiatives unlock new rewards, artifacts, and quests in games. The more the earth heals, the more engaging the games become, the more people play. The more people play, the more revenue is generated, the faster the earth heals, and so on.
We can use digital twins and game engine APIs to build explorable digital representations of physical spaces that respond, in real-time, to changes at their sources. One member of the UN Environmental Programme’s Playing for the Planet initiative, Kenya-based game company Internet of Elephants, has developed a series of games that incorporate wildlife camera trap data and other data about wildlife migration patterns. A July 2022 Stanford Social Innovation Review paper, “Using the Metaverse to Connect and Protect the Natural World,” (Shaw, Boudinot) describes how such data can inform wildlife conservation efforts.
The WIL represents a turn away from an over-reliance on human conscience as our primary means to confront the climate crisis. It embeds within our digital systems algorithmic mechanisms that respond to environmental crises by design, without depending on the weak link of human agency, which has proven to be susceptible to misinformation, propaganda, overstimulation, greenwashing, and an over-emphasis on individual consumer choices (i.e. one’s “carbon footprint,” a term coined by British Petroleum and PR firm Ogilvy Meyer). The WIL produces positive climate action as a byproduct of its use, wholly independent of human intent. A further unconscious but essential byproduct of the WIL is the habit of time-perspective flipping, which we believe engenders deep understanding of our own meta-prompts and faster acceptance of and navigation with the superdisruptors that we ourselves have created.
Technologies necessary to build and manage the WIL
The WIL isn’t one particular technology or platform. It is a protocol developed to orchestrate technologies and platforms. Technical elements of the WIL include:
- Data gathering devices that include IoT monitors, LiDAR, GIS, mycelium network monitors, satellite imagery, drones, citizen scientist reporting, et al.
- Public and private clouds such Google Cloud Platform, Microsoft Azure, and Amazon AWS, as well as hybrid, multicloud, and on-premises platforms.
- Data lakes, data warehouses, data lakehouses for managing structured and unstructured data.
- Quantum computing.
- Cloud security platforms in the style of Mandiant and Google Chronicle, configured to proactively alert users to emergencies reflected in environmental data (as opposed to cybersecurity threats).
- Public and private data repositories, including Google Earth Engine, Microsoft’s Planetary Computer, NOAA, and ESRI’s ArcGIS.
- 3D Modeling programs including Unity and Unreal Engine, digital twins, and space-time cubes.
- Game platforms including PC, Xbox, Playstation, and mobile.
- Generative AI, including ChatGPT, Copilot, Gemini, Vertex, Midjourney, et al.
- Blockchain for credit assignment, geographical locations, and biotags.
All the technologies necessary to build a platform that can monitor the earth, identify environmental crises, and automatically direct resources to confront these crises currently exist. Our task is merely to orchestrate them. Artificial intelligence comes into play at two stages. First, as an assistant to code the various systems and platforms that comprise the WIL, and secondarily as a middle-agent monitor to provide feedback to the system and the real world.
The WIL from the Point of View of Data
Imagine a mountainside in the Cascade range of the Pacific Northwest, dense with Douglas fir trees, that has been ravaged by wildfire. One way to remedy such a tragic turn of events is through drone-based reforestation. Today, companies like Toronto’s Flash Forest and Seattle-based Mast Reforestation employ fleets of drones that shoot seed pellets into the soil to help mitigate against wildfires exacerbated by climatic heating. In addition to the ease with which drones can access previously hard to reach areas, the geographic coordinates of each planting can be recorded and saved in the cloud.
Let’s imagine our hypothetical mountainside as the site for such a reforestation project. As each seedling takes root in the physical landscape, the data representing that seedling takes root in the cloud. Envision a single tree on this mountainside as it begins a double life, both as a botanical entity living on earth, and as a representation existing in a digital platform.
Routine drone and satellite imaging scans of the landscape determine that this particular tree has begun to grow and thrive. A digital twin of this parcel of land provides reforestation experts with a 3D representation of the tree that they can access on screens or with virtual reality goggles, anywhere in the world.
To measure the health of this forest over time, researchers create a Space Time Cube, a temporal-geographical model that depicts changes in physical landscapes. As part of a US Forestry Services program, the public is invited to visit the digital representations of this forest at any time, on their devices, from the comfort of their homes, as part of a digital park pass program.
The digital twin of the forest serves as the basis for an adventure game. Over time, rewards in the game in the form of achievements, skills, and digital goods are tied to fluctuations in the data. As more trees are planted, more treasures are unlocked in the game. As players spend real money on the game itself, a portion of the proceeds are directed toward boots-on-the-ground reforestation initiatives.
Our once-wildfire-scarred landscape has begun once again to teem with life. Camera trap data records the reappearance of deer, owls, squirrels, and other animals, and this data, too, is uploaded to the cloud via remote sensing devices, and used to trigger changes in the landscape’s digital twin.
Meanwhile, monitors embedded in the soil record the electric pulses traveling through fungal mycelium networks, the communications infrastructure of the Wood Wide Web. This data, too, is uploaded to the cloud, where AI translates the language of trees into human languages. Soon we’re able to determine whether the forest is fighting off an infestation of a particular kind of beetle, or whether certain mineral resources in the soil have become scarce. In a sense, we begin to meet the plants with whom we share the planet in a noospherically mediated platform, where we engage in more nuanced communications than have been possible in the physical world.
As more data about the earth streams into the cloud, AI/ML algorithms detect patterns and generate intelligence about the health of the mountainside, which allow humans working in the field to take specific action and work in greater harmony with the forest. Humans thus become a more deeply entwined component of the forest’s immune system, which now extends from the soil into server farms and wireless signals.
As the forest thrives, the digital twin thrives, too, and metrics related to the health of the forest trigger greater digital rewards for the humans who work to keep it healthy. In every biome on earth–from coral reefs to deserts, plains, and rainforests–the World Integration Loop provides a feedback mechanism through which positive planetary data triggers positive economic returns for humans, as natural resources become more valuable for their capacity to generate data than for providing material in an extractive economy.
Final Thoughts
Going back to the ideas of superdisruptors being agents of change that induce movement toward a meta-prompt, thinking in a linear unfolding perspective, what meta-prompt could be created by the WIL? Thinking in a bird’s-eye-view perspective, what meta-prompt could necessitate the emergence of a bidirectional causal engine like the WIL? We think it is the same meta-prompt, something like, “Please move toward a more holistic understanding of humanity’s interdependence with each other and the planet, through all of time.”
By following this meta-prompt, the planet as a whole comes to know itself in greater detail. One might imagine that the entire planet, its atmosphere, and all of its organisms attain a degree of consciousness about our own natural processes that has heretofore proven elusive. In this line of thinking, if something like the WIL helped move us in this direction, can this intelligence truly be called artificial?
