This morning, over coffee, I found myself reading a white paper from the Collective Intelligence Project—a research organization proposing new governance models for transformative technologies like artificial intelligence, biotechnology, and geoengineering. Their framing is sharp: we face a “transformative technology trilemma” forcing us to sacrifice one of three critical values—progress, safety, or participation. They argue we need a fourth path, building collective intelligence systems that encompass all three goals.

It’s intellectually compelling work. Their advisory board is impressive. The problems they’re addressing are real and urgent. And yet, as someone who has spent 36 years watching technology waves reshape ecological research—from GIS to sensor networks to AI—I find myself asking a different question: Will this become infrastructure, or will it remain noise?

The Pattern Recognition Lens

When you’ve been in field ecology long enough, you develop pattern recognition. You learn to distinguish signal from noise, trend from fluctuation, reversible from catastrophic. I’ve watched technology adoption cycles from the inside—not as a consumer or theorist, but as someone deploying systems in remote locations, building networks across research stations, integrating new capabilities into ongoing projects.

Some technologies stuck. Geographic Information Systems became unavoidable once you realized you could overlay spatial data in ways that hand-drawn acetate sheets never allowed. Databases replaced card catalogs because queryable storage solved fundamental problems. The internet connected isolated information sources with network effects that made adoption inevitable. Digital imaging replaced film for reasons of cost, speed, and manipulability that transformed field ecology. Mobile computing enabled sensor networks in locations that were previously unreachable.

These weren’t just “better tools.” They became infrastructure because they had forcing functions. Researchers needed them regardless of institutional politics. The work became impossible without them.

The Collective Intelligence Question

The term “collective intelligence” isn’t new, though it sounds like Silicon Valley rebranding. The concept traces back to Condorcet’s 1785 jury theorem—the mathematical observation that if individual voters are more likely than not to be correct, the probability of a correct group decision increases with group size. Émile Durkheim in 1912 argued that society itself constitutes a higher intelligence transcending individuals. Mid-20th century, Douglas Engelbart linked collective intelligence to organizational effectiveness, predicting that augmenting human intellect would yield multiplier effects in group problem-solving.

Pierre Lévy coined the modern usage in 1994, describing the internet as enabling “universally distributed intelligence.” Thomas Malone founded MIT’s Center for Collective Intelligence in 2006 to study “how people and computers can be connected so that—collectively—they act more intelligently than any person, group or computer has ever done before.” Woolley and colleagues identified a statistical “c-factor” for group intelligence in 2010, though subsequent meta-analyses found only moderate correlations with actual performance.

So there’s legitimate academic heritage here—not just buzzwords. But the field is fragmented. Different disciplines study group-level intelligence using different methods, asking different questions, rarely synthesizing across boundaries. The Collective Intelligence Project is using it in the governance and institutional design sense—how groups make decisions about technology development. This is closer to Lévy’s philosophical vision than to empirical psychometrics.

Where the Metaphor Breaks Down

One CIP project caught my attention: partnering with the Earth Species Project to explore AI-assisted animal communication. They surveyed over a thousand people across 67 countries and found that 60% believe animals should participate in human democracy in some capacity—through voting, advisory roles, or representation.

My first thought was of Gary Larson cartoons. “What we say to dogs / What they hear: blah blah GINGER blah blah blah blah.” The humor comes from the disconnect between human interpretation and animal experience. CIP’s “democratic participation” of animals makes the same category error—it assumes translation into human political structures is the goal.

But beneath the questionable framing lies legitimate science. Pattern recognition in animal vocalizations correlated with observed behaviors is valuable observational data. Understanding that sperm whales have dialects, that prairie dogs have alarm calls encoding predator information, that bees have directional communication—this extends our observational capacity. It’s another sensing modality for assessing ecosystem health.

That’s different from “asking whales to vote.” Animals are already communicating about ecosystem conditions. The question is whether humans build observation infrastructure to detect and interpret those signals as data, not as opinions to be incorporated into voting mechanisms. Whales don’t need democratic representation in shipping lane decisions—but distributed hydrophone networks with pattern recognition of distress vocalizations would provide real-time feedback about anthropogenic ocean noise impacts. That’s not democracy. That’s instrumentation.

This distinction matters because it reveals what CIP is actually proposing: deliberation mechanisms without the observation infrastructure. They’re designing sophisticated processes for aggregating preferences and making choices—liquid democracy, quadratic voting, citizens assemblies. These are tools for the decision layer. But without the perception layer, what are you deciding about?

The Forcing Function Problem

My Macroscope work has always recognized you need the sensing infrastructure first. The EARTH domain monitors abiotic conditions. The LIFE domain tracks species and biodiversity. The HOME domain observes immediate environments. The SELF domain captures personal metrics. That’s the perception layer—distributed observation creating shared situational awareness. Only then can you make informed decisions.

CIP is working entirely on the decision layer. And that creates a vulnerability: process innovation without forcing functions rarely becomes infrastructure. Sophisticated deliberation mechanisms require voluntary adoption. They need participants who believe governance is meant to solve problems rather than reward allies and punish enemies. They assume prerequisites—good faith engagement, willingness to engage with complexity, institutions capable of implementing decisions, basic agreement that expertise matters.

Here in the western states, we’re watching those prerequisites crumble. The federal programs that supported long-term ecological observation, the NSF centers I worked with, the institutional continuity that made multi-decade research possible—all built on assumptions that some things transcend political cycles. That assumption is shattering, not gradually eroding.

CIP’s white paper describes three failure modes for governing transformative technology: Capitalist Acceleration (sacrificing safety for progress), Authoritarian Technocracy (sacrificing participation for safety), and Shared Stagnation (sacrificing progress for participation). They propose a fourth path through collective intelligence R&D.

But there’s an unacknowledged fourth failure mode we’re experiencing in real time: deliberate institutional destruction. What happens when some actors benefit from preventing coordination? When dysfunction becomes strategy rather than problem? When the goal is dismantling the institutions that would implement whatever gets decided?

You can design all the liquid democracy mechanisms you want. They’re irrelevant if one faction’s aim is to dismantle the institutions that would implement them.

Earthquake Preparedness

Growing up in Los Angeles, I experienced devastating earthquakes—Sylmar, Loma Prieta, hundreds of smaller events. Not once did these compel me to leave California, but they created a background awareness of what to do when it happens. You retrofit buildings not because you believe you can prevent earthquakes, but because systems that flex survive better than systems that shatter.

The political tsunami sirens are activated here on the west coast. Oregon, Washington, California are preparing for federal program failures the way my research stations got earthquake-proofed. State environmental agencies figuring out how to maintain monitoring networks if EPA collapses. Universities preparing to fund research if NSF gets gutted. Regional compacts to share data if federal coordination disappears.

The earthquake metaphor extends only so far. Geology doesn’t have agency. The San Andreas fault isn’t trying to destroy things. This political moment is different—the destruction is intentional. That changes the psychological calculus in ways that earthquake preparedness doesn’t quite capture.

And yet the response remains appropriate: You retrofit. You prepare. You maintain what you can. You don’t abandon your position. You watch for warning signs, have protocols ready, and continue the work. My Macroscope project doesn’t stop because federal institutions are failing. Sensor networks keep sensing. Species keep existing. Ecological baselines keep shifting regardless of political chaos. The work remains.

What Sticks and What Doesn’t

So where does this leave the Collective Intelligence Project’s vision? I won’t judge that they aren’t doing relevant work—thinking people searching for solutions is always important. But my pattern recognition tells me to watch for forcing functions.

The technologies that became infrastructure had compelling reasons for adoption independent of politics or ideology. Researchers needed databases whether liberal or conservative. GIS worked regardless of who controlled Congress. Network effects transcended partisan identity.

Collective intelligence governance mechanisms lack that independence. They’re processes that require participants to choose wisdom over convenience, long-term thinking over immediate advantage, collective benefit over tribal competition. That’s aspirational, not inevitable.

The animal communication science might become infrastructure—bioacoustic monitoring for conservation has forcing functions in biodiversity loss and ecosystem change. The “animal democracy” framing won’t stick—it’s anthropomorphic projection without practical application.

CIP’s deliberation tools might get picked up by states, cities, or other jurisdictions trying to maintain institutional sanity. Some mechanisms might prove useful in rebuilding collective decision-making capacity after the current crisis passes. Or perhaps the conditions for that kind of work are evaporating in real time.

I don’t know which. But I know what I’m doing: maintaining observation infrastructure, keeping sensors running, documenting what’s happening. Building systems that flex rather than shatter. Preparing without certainty, building resilience without guarantees.

That’s what you can do when you can’t prevent the earthquake but also can’t evacuate. You retrofit what you can, maintain your networks, and continue the work. The baseline is shifting—in ecosystems, in institutions, in civic ecology. Pattern recognition doesn’t tell you everything will be fine. It just tells you how to keep observing while things change around you.

The Collective Intelligence Project is proposing sophisticated tools for a future we hope to build. I’m maintaining the observation infrastructure we’ll need whether that future arrives or not. Maybe both are necessary. Maybe only one will matter. But the sensors keep sensing either way, and the data keeps accumulating, and that’s the work that persists when institutions fracture.

Some morning, perhaps, we’ll look back and recognize which parts became infrastructure and which parts were noise. Until then, we prepare, we adapt, and we keep the baseline measurements running. That’s what 36 years of pattern recognition teaches: the work of observation continues, regardless.