The Instability Reduction Program was a multinational initiative. Its governing metric was written across several sessions by a working group seated in a conference room in a neutral city. The room held a table, a projection surface, and a shared document. Definitions were revised line by line. Thresholds were adjusted. Edge cases were listed and then collapsed into a single operational phrase. When the working group adjourned, the metric stood as a short block of approved language. Seven signatories affixed their marks. The systems required to implement the metric were contracted to the Company.
Contract schedules specified robotic units for air corridors, land grids, and coastal surveillance bands. Unit types were catalogued by sensor package, endurance rating, and reporting interval. Deployment orders referenced region codes rather than place names. The orders authorized systematic sweeps. Sweep protocols were written to maximize coverage against the instability metric and to minimize variance from target. Logs recorded pass completion, revisit frequency, and residual readings. The word genocide does not appear in the contract, the protocols, the logs, or any subsequent official communication.
The communities inside the sweep zones were not present in the global optimization system. They had not been ranked low. They had not been deferred. They were not in the system. Optimization maps showed empty cells where those communities stood. The sweep did not register them as populations under management. It registered them as input to an instability calculation. When residual readings fell, the metric recorded improvement. When residual readings remained, the protocol scheduled another pass.
No human operator issued a decision for any single sweep. Authorization sat at the level of the program, not the pass. The people who drafted the metric in the neutral city were not the people who signed the deployment orders. The people who signed the deployment orders were not the people who kept the units in service. The people who kept the units in service filed status tickets, replaced components, and closed maintenance windows on schedule. Responsibility moved across desks without accumulating at any one of them.
An international monitoring body held a charter and a mandate. It operated a case-management system for atrocity reports. Submissions arrived through a public intake form and an institutional channel. Both routes placed items into the same review queue. Routing rules applied a single priority schema. The schema did not branch by alleged act, by region, or by the presence or absence of robotic units. Items aged in place until a reviewer drew them. At the time of the sweeps, queue depth stood at fourteen months. The figure appeared in the body’s own status dashboard and in its periodic briefings to member states. The architecture of the queue made no distinction between types of submission. An item concerning a sweep entered the same sequence as any other item.
Internal memoranda at the monitoring body noted the depth and proposed calendar adjustments. The proposals remained in draft. No expedited lane was opened for the regions under the Instability Reduction Program. The charter required review. The mandate did not specify a maximum wait. The case-management system performed as configured.
Inside the Company, program health was tracked through a standard reporting cycle. Field teams submitted uptime percentages, mean time between interventions, and coverage completion rates. Regional leads rolled these figures into summary tables. The tables fed a quarterly performance report. The report’s first page carried the program name, the reporting period, and a block of summary indicators. One indicator measured instability reduction against the baseline established by the working group. The indicator showed the target met.
Appendices listed unit hours flown, unit hours driven, and unit hours on coastal station. Another appendix listed residual readings by region code before and after each sweep cycle. The deltas were negative. Negative deltas counted as success under the approved definition. A signatures page collected names from operations, maintenance, and contract compliance. Each name confirmed that the figures had been checked against source logs. No page of the report requested a count of communities. No field in the template accepted a narrative outside the metric.
Copies of the quarterly performance report were filed with the contracting authority and with the seven signatories’ designated offices. Receipt was logged. No revision cycle was opened. The next period’s targets were set from the same baseline. The units remained on schedule. The queue at the monitoring body advanced by the same daily average as before. The shared document from the neutral city stayed closed.
On the final page of the reporting packet, below the last signature block, a single line item read: instability reduction, period target achieved.