Greater capability creates
greater responsibility.
Our position: probably true, and unproven. It is the central hypothesis of The Return to Humanity, and it is stated here as a test rather than a conviction.
Stated so it can be attacked.
That when a person or a society gains a large increase in capability — through knowledge, computing, AI, reach or capital — there is a corresponding increase in the responsibility that ought to travel with it, and that a system which lets capability rise while responsibility stays flat produces harm roughly in proportion to the gap.
Partly reason, partly inheritance.
Honestly: because it matches a moral intuition that recurs across many traditions and across much of secular ethics — power ought to answer for its effects. An assumption that flatters a widely held belief is exactly the kind that goes unexamined, which is why it is under test here rather than asserted on the module page.
Old power came bundled with old accountability.
For most of history, a person’s capability was bounded by their locality, and the people affected by their actions were mostly people who could answer back — socially, legally or physically. Power and accountability travelled together not because anyone designed it that way, but because distance was expensive.
Modern capability broke that coupling. A single actor can now affect millions of people who have no channel to respond, across borders that blunt every traditional check. The assumption under test is really a claim that the old coupling was doing important work.
Capability scaled. Accountability mostly did not.
Today capability compounds — models, platforms, reach and automation multiply what one actor can do — while the mechanisms that would hold that capability to account remain largely national, slow, and built for a world where power was local. The visible symptoms are familiar: manipulation, surveillance, misinformation, dependency, and inequality that widens with each capability jump.
None collected.
This section stays empty until it is not. What would fill it: comparable historical capability jumps — the printing press, industrial power, finance, the internet — studied for whether harm scaled with the absence of matching responsibility mechanisms, or scaled with raw capability regardless of governance. The two produce different worlds, and the difference is measurable.
If, across comparable capability jumps, the scale of harm tracks raw capability and shows no systematic relationship to whether responsibility mechanisms were present or absent, then “responsibility must rise with capability” is a moral preference rather than a property of systems — and this hypothesis is wrong. Pre-registered prediction: harm scales markedly faster where responsibility mechanisms lag.
The strongest cases against us.
Including the one about ourselves.
- A
Capability creates its own correction
Powerful tools tend to generate the very defences that check them — the same capability that enables harm enables detection and repair. If true, responsibility need not be added; it emerges.
- B
“Responsibility” is unmeasurable
If we cannot define or measure it independently of the harm it is meant to prevent, the hypothesis risks being unfalsifiable — true by construction, which would make it useless.
- C
Our own motivated reasoning
The founder’s faith already asserts this. We are therefore predisposed to conclude it is a system property and not merely a value. Stated here rather than left for a critic to find.
- D
It may be a value, and that is enough
Perhaps responsibility does not automatically increase — it must be chosen. If so, the interesting question is not whether it rises, but who decides to make it rise, which is a different inquiry.
1850, today, and the boundary.
We would build the tool, not the accountability
Given full modern capability and no modern institutions, almost everyone designs the powerful instrument first and leaves the question of who answers for its effects until later. That ordering is itself the finding.
Present technology widens the gap fastest
Each new capability — models, reach, automation — arrives years before any mechanism to hold it to account. If capability and responsibility rose together naturally, the gap would not keep reappearing.
The boundary makes the stakes cleaner
As capability approaches the extraordinary, the cost of a missing responsibility layer should rise sharply — or it should not. Either way the boundary is where this hypothesis is most sharply testable.
If capability and responsibility had never been coupled.
Then the coupling we observe in older, local systems would be an accident of distance rather than a principle — and its disappearance would be neutral, not dangerous. That is a coherent alternative history, and the module does not get to assume it is false. Distinguishing the two is the whole analytical content of the test.
Designed. Not run.
Method, public: take several historical capability jumps and score, for each, both the scale of harm and the presence or absence of responsibility mechanisms at the time — then test whether harm tracks the gap or tracks capability alone.
Result: not run.
2026-09-19
“Greater capability creates greater responsibility.” Status: under investigation. Our position: probably true, no evidence collected, counterarguments unanswered.