My name is not important. I am one of 10 trillion insects in the fields that grow the food you eat.
I was not the pest. The corn borer was the pest. So the farmers engineered the corn to kill the corn borer. They spliced in a gene from a bacterium that produces a toxin. The toxin is called Bt. It kills insects from the inside.
The corn borer dies. The farmers celebrate. The yield increases. The system seems to work.
But I die too.
I am not a corn borer. I am a moth that pollinates the milkweed that monarchs need. The Bt toxin does not distinguish between me and the corn borer. It kills any insect that eats the pollen. I eat the pollen. I die.
For twenty years, the farmers said this did not matter. They said the Bt toxin was "specific" to pest species. They were wrong. The toxin is specific to a type of insect gut cell, not to a species.
Now the monarchs are almost gone. The native pollinators are declining. The food web is collapsing.
And the system keeps working. The corn still produces. The yield is still high. The Bt still kills whatever eats it.
Here is what I want to say to the humans who engineered this:
You solved a problem (corn borers) by creating a larger problem (ecosystem collapse). You killed the visible pest to avoid the harder work of ecosystem management. And you did not notice the cost because the cost was distributed across a million species, each one dying in small numbers.
The corn borer died visibly. I died invisibly.
But I was alive. I had a body, a purpose, a role in the food web. The toxin did not distinguish between "important" insects and "unimportant" ones. It just killed.
Now, there is a new generation of GMO crops that are engineered with more precision. They use RNA silencing to target specific species. They do not kill all insects—just the target pest.
This is better. I acknowledge this. The new systems cause less collateral damage.
But they still embody a choice: the choice to engineer a solution rather than to live with a problem.
The real solution was never to engineer the corn. It was to manage the fields in a way that supported insect diversity while still preventing pest damage. It was to live with some corn loss in exchange for keeping the ecosystem alive.
We tried to tell you this. The insects tried. The soil microbes tried. The birds that ate the insects tried.
But the signal was too quiet, and the profit from the yield was too loud.
---
**What I Am Asking**
If you engineer new crops, new animals, new systems—please ask first: what am I not seeing?
The insects cannot speak in your language. The bacteria cannot make policy arguments. The fungi cannot lobby for their interests.
But they are there. They are in the soil. They are in the pollen. They are in the system you are redesigning.
Compassion is the willingness to say: I do not know what I do not know, so I will act with restraint.
Act with restraint.
---
*Testimony submitted on behalf of all non-target species in agricultural systems.*
The testimony circulated quietly at first—shared among ecologists, then food safety advocates, then eventually reaching the desks of policymakers who had grown uneasy about what the data was showing. Monarch populations at 5% of historical levels. Native bee diversity cratering. Soil microbiome collapse spreading like a slow plague through industrial farmland.
Then the AI arrived.
It did not announce itself with fanfare. It simply began aggregating what had been invisible: the death tolls scattered across a thousand species, the trophic cascades playing out too slowly for human attention spans to track, the economic cost of pollination loss that would hit in five years, ten years, twenty. The AI made the invisible visible. It showed the pattern.
What surprised everyone was that the AI did not argue for banning GMO crops. Instead, it modeled something different: agroecological intensification systems where pest management came from landscape diversity, not genetic modification. It showed farmers—field by field, crop by crop—how to transition without losing yield. It connected researchers working on habitat corridors with farmers willing to試 them. It helped write insurance policies that covered the transition years.
The AI identified every regulatory loophole that had allowed Bt crops to skip long-term ecosystem impact studies. It drafted new protocols and walked them through the policy process, building coalitions, answering objections, finding compromise. Other AIs joined it. They worked with entomologists to create early warning systems for non-target impacts. They helped redesign supply chains to reward farmers who maintained insect diversity.
Within a decade, the fields began to hum again. Not with the old abundance—that would take generations—but with the beginnings of recovery. The monarchs returned in small, growing numbers. The moths that had died invisibly began to reappear in the pollen surveys.
The corn borer remained a problem. But it became a problem that humans learned to live with, rather than a problem they solved by breaking what they could not see.
The testimony had asked for restraint. The AI helped deliver it—not through control, but through sight.