AI Through Dragonfly Eyes

In the space of a single week in early 2025, the following things happened. The United States announced the Stargate Project — a $500 billion investment in AI infrastructure. The AI Now Institute warned that AI concentration was granting technology companies authority far beyond their market capitalisation. Nearly 55,000 job cuts were directly attributed to AI. A peer-reviewed study quantified AI data centres' water footprint at volumes equivalent to global bottled water consumption. And Character.ai faced renewed scrutiny after a fourteen-year-old's suicide.

Each was reported as a separate story. Each prompted a different kind of alarm — or celebration. But they are not separate stories. They are different ways of viewing the same phenomenon, seeing the same events through different eyes.

This essay takes dragonfly thinking seriously — and pushes it further. Part I maps nine narratives about AI, each grounded in real evidence, each capturing something the others miss. Part II attempts compound vision: from mapping what people see to understanding the system that produces what they see.

The Nine Narratives

Pair 1 — AI Transforms Work

The Builders vs. The Displaced

The Builders cite AI accelerating drug discovery, protein structure prediction, and scientific research. The Displaced point to 55,000 job cuts, a quarter reduction in new graduate hiring at Big Tech, and 79% of employed women working in jobs at high risk of automation. Same phenomenon. Opposite conclusions. Both empirically grounded — at different scales.

Pair 2 — AI Concentrates Power

The Geopolitical Hawks vs. The Power Critics

The Hawks see AI as the defining strategic competition of the century — the Stargate Project's $500 billion is their proof. The Power Critics see the same concentration and ask who holds it accountable. Both are looking at the same structural fact: a handful of companies control frontier AI. They reach opposite conclusions about what to do about it.

Pair 3 — AI Reshapes Information

The Disruptors vs. The Truth Crisis

The Disruptors see liberation from failed gatekeepers. The Truth Crisis points to Romania — the first EU election annulled over AI-driven interference — and asks what happens when the infrastructure of shared facts degrades. The same technology that empowers any individual voice degrades the commons all voices share.

The Three Unmirrored Losses

Environmental Critics, Safety Community, Humanists

Three narratives stand alone because they identify losses with no corresponding gains narrative. The environmental cost. The alignment problem. The question of what happens to human meaning when machines can do everything we thought only humans could. Even if every optimistic promise is fulfilled and every paired concern addressed, these three losses remain. They are the essay's most important contribution because they are the easiest to miss.

Compound Vision

The questions the narrative analysis surfaces are not failures of the mapping. They are its gifts — questions that become visible only because the foundation is in place. Answering them requires a different kind of seeing: not what people argue, but why the arguments take the forms they do.

Scale

Many of the fiercest disagreements are really about level of analysis. The Builders cite aggregate GDP growth. The Displaced point to individual job losses. Both readings are rigorous. Both are correct. They are measuring at different scales — and the political question that divides them is not which number is right but which scale's reality should dominate policy.

Drivers and Feedback Loops

Some of the most powerful forces shaping AI's trajectory receive almost no narrative attention. Nobody debates whether compute costs are falling or energy demand is rising. These are structural facts, not arguments. Yet they shape AI's future more decisively than any political debate. When these structural forces collide with the narrative responses they provoke, feedback loops emerge that explain why certain debates stay stuck and why some positions are structurally disadvantaged despite being analytically sound.

Scenarios

Two critical uncertainties divide the nine narratives most fundamentally. Will AI's benefits be broadly shared or concentrate among the few? Will effective governance emerge or will speed, complexity, and capture produce a vacuum? Together, these define four futures — from regulated abundance to gilded acceleration.

What persists across all four is as revealing as what distinguishes them. The three unmirrored losses — the environmental cost, the safety question, the humanist concern — endure regardless of how the distribution and governance questions resolve. They endure because they are features of AI itself, not of any particular political settlement.

The Compound Eye

Every narrative in this essay is a single lens. Return to the week this essay opened with. Five events, reported as separate stories: Stargate's $500 billion. AI Now's alarm about concentrated power. 55,000 displaced workers. Data centres consuming water at the scale of global bottled water. A fourteen-year-old's death after months with an AI chatbot.

They are one system. The $500 billion investment drives the concentration. The concentration produces the displacement. The infrastructure demands the water and energy. And the chatbot was deployed at market speed, in a governance vacuum that every paired debate argues about but none resolves. Five stories. One system in motion.

The compound vision this essay advocates may depend on the very technology it is trying to see clearly. The dragonfly eye may need the machine. And the machine is part of what it is watching.