BCT
Benefit-Cost Transfer
Did the protected benefit depend materially on shifting cost, risk, uncertainty, or correction work outward?
The Full Ledger
Agency, Power, Cost, and Responsibility in the Age of AI
A framework for tracing how decisions distribute benefit, burden, authority, and responsibility as AI extends the reach of human and institutional action.
Premise
AI is rapidly expanding the ability of people, organizations, and autonomous systems to act at machine speed and unprecedented scale.
Throughout my career, I have seen how decisions continue through requirements, software, controls, products, and institutions long after the original decision-makers have moved on. AI intensifies that pattern. It can amplify good judgment, but it can also multiply narrow objectives, hidden assumptions, externalized costs, and gaps in responsibility before the consequences are fully understood.
I wrote The Full Ledger because our power to act is growing faster than our ability to account for the results. Its purpose is to help keep agency, power, cost, and responsibility connected as AI increases the speed, reach, and magnitude of human and institutional action.
A method for tracing responsibility
A system can be tightly closed around its assigned objective while remaining open around the burdens it creates. The framework asks whether evidence from the people and systems carrying those burdens can reach responsible authority and change what happens next.
The responsibility loop
- 01Objective
- 02Action
- 03Benefit
- 04Burden
- 05Feedback
- 06Authority
- 07Repair
- 08Changed next decision
The loop closes only when the evidence can change the next decision.
Three bounded questions
ECE
Externalized Cost Exposure
Who or what was exposed, and were those burdens represented in the decision?
UR
Unresolved Responsibility
After credible notice, what appropriate and feasible responsibility remained unfulfilled?
Together, BCT, ECE, and UR support the FLA Gap Index, a sortable investigative-priority signal. Its meaning depends on the evidence record and case boundary.
The book
The narrative argument, working model, demonstrations, and executable framework are contained in one volume.
Contents+
Part I
The Argument
The Introduction, ten chapters, and Conclusion make the transition into AI agency understandable at a human scale.
Part II
The Working Model
The Responsibility Calculus preserves definitions, equations, scoring rules, uncertainty, and worked examples.
Part III
Five Demonstrations
Public-record cases show how the method operates, where interpretation enters, and how corrective evidence changes the result.
Part IV
The Executable Framework
The canonical agent, glossary, sources, and analytical index allow readers to inspect, test, reproduce, challenge, or adapt the method.
Scope and boundary+
- A complete first volume, Book Revision 12.2.3.
- A working Responsibility Calculus and executable agent specification.
- Five bounded public-record demonstrations with evidence cutoffs, countercases, uncertainty, and corrective records.