Chapter 25 - THE ELECTION MODEL NO ONE VOTED FOR
Meridian’s restricted server was called CIVIC RESPONSE.
The name sounded harmless.
So had every other system before it.
Claire accessed the dashboard using voting-control credentials that should never have belonged to her.
A timer appeared.
LEGAL STATUS REVIEW IN PROGRESS.
Someone had already challenged the transfer.
Marcus told them to move quickly.
The first folders contained familiar material.
Pensions.
Hospital payment behavior.
Housing instability.
Labor disputes.
Then Claire opened CIVIC.
The data categories changed.
County.
Precinct.
Household.
Employment sector.
Medical debt.
Mortgage distress.
Public-benefit reliance.
Union membership probability.
Church attendance estimates.
Veteran status.
Caregiving burden.
Educational debt.
Not individual voting records.
Those were secret.
Instead, Meridian modeled political participation.
Who was likely to vote.
Who might stay home.
What type of economic disruption reduced turnout.
Elena felt cold.
“This is campaign software.”
Marcus answered through speakerphone.
“No. Campaign software tries to persuade voters.”
“What does this do?”
“It models friction.”
The distinction was horrifying.
Phase Four was not primarily designed to change political opinions.
It predicted how administrative inconvenience changed participation.
A delayed paycheck.
A shift schedule changed at the last minute.
A bus route reduced.
A pharmacy closure.
A child-care shortage.
A benefits review.
None of those events told someone how to vote.
They could determine whether someone had time to vote at all.
Ruth understood instantly.
“You don’t have to change somebody’s mind if you can make Tuesday impossible.”
Marcus said that was exactly the concern that made him oppose the project.
Meridian originally developed CIVIC RESPONSE for emergency planning.
Governments needed to estimate whether hurricanes, transit failures, or school closures would disrupt public participation.
Edward’s team discovered another application.
Targeted friction.
Small administrative decisions could change turnout without obvious election interference.
No fake ballots.
No hacked machines.
No direct intimidation.
Just unequal inconvenience.
Naomi asked whether the system had ever been used.
Marcus hesitated.
“Yes.”
The first tests were not elections.
Union votes.
Tenant association elections.
Pension trustee elections.
Professional board elections.
Bellmont’s fake employee vote had been crude by comparison.
Phase Four learned which obstacles suppressed participation among specific groups.
Then it expanded.
Claire opened a folder marked PILOT MUNICIPAL.
Three cities.
Local elections.
The software had recommended operational changes.
One county delayed public-employee overtime payments during early voting.
Another changed bus maintenance schedules.
A third accelerated benefits recertification notices during the final week of registration.
Were officials told these actions would affect turnout?
The server contained no explicit command.
Only predicted impact.
Policy decision.
Operational benefit.
Participation consequence.
Everything separated.
Responsibility dissolved across departments.
Elena asked Marcus:
“Who bought it?”
“Not campaigns.”
“Who?”
“Consultants. contractors. public-private partnerships.”
Organizations with plausible nonpolitical reasons to alter operations.
The system created influence without campaign fingerprints.
Then Ruth saw a national map.
HOUSEHOLD FRICTION INDEX.
Millions of points.
Red.
Yellow.
Green.
The Bellmont Phase Four household list was only a sample.
The national database was far larger.
“How did they get this data?” Ruth asked.
Marcus answered:
“Brokers.”
Credit data.
Commercial consumer files.
Public records.
Health-related purchasing signals.
Employment information.
Not necessarily illegal individually.
Combined, devastating.
The software identified households where one additional administrative burden could suppress civic participation.
Elena remembered Catherine Vale.
PHASE FOUR REQUIRES NATIONAL CONSENT.
They had interpreted it as congressional authorization.
Maybe wrong.
National consent did not mean a vote approving Phase Four.
It meant ordinary citizens gradually consenting to institutions collecting enough data to model them.
Terms of service.
Credit applications.
Benefit forms.
Employment portals.
Healthcare systems.
Consent scattered into hundreds of clicks.
The vote America never cast.
Then Claire opened CURRENT DEPLOYMENTS.
One program was active.
2026 NATIONAL RESILIENCE STUDY.
No public description.
Sponsor:
American Civic Continuity Consortium.
Members included Meridian.
Mercer National.
Two hospital networks.
A logistics company.
Three data brokers.
And the United States Pension Stabilization Board.
Catherine’s board.
The consortium claimed to study whether economic shocks could disrupt public participation during emergencies.
The next scheduled test:
November.
Not a federal election year.
Several states had major gubernatorial, judicial, and ballot-measure elections.
One pilot state stood out.
Pennsylvania.
Why?
Large urban workforce.
Rural hospitals.
Union density variation.
Significant medical debt.
Transit differences.
An ideal laboratory.
The system listed 240,000 households.
Ruth searched her name.
Not included.
Claire searched herself.
No.
Then Naomi searched Bellmont workers.
Thousands appeared.
The cooperative had become a data source.
Their hardship patterns were being used to calibrate voter-friction models.
Workers who thought they had won Phase Three had unknowingly trained Phase Four.
Ruth’s expression changed.
“They didn’t need Bellmont to survive.”
“No,” Elena said.
“They needed the data.”
That was Edward’s prediction.
Workers had to win.
The cooperative victory produced authentic information about solidarity.
Who gave money.
Who attended meetings.
Who refused settlements.
Who changed votes when medical benefits disappeared.
A resistance dataset.
Then Claire found another metric.
TRUSTED MESSENGER INDEX.
Phase Four mapped who communities believed.
Pastors.
Nurses.
Teachers.
Local organizers.
Union stewards.
Parents.
Ruth Carter’s score was exceptionally high.
That explained why she had been targeted.
Not because she controlled money.
Because she influenced whether others believed institutions.
If Phase Four wanted to suppress participation quietly, trusted local people could expose the pattern.
They needed to be discredited first.
Ruth was not a collateral target.
She was a civic threat.
Then Elena searched her own name.
ELENA HALE.
Trusted messenger score: 6.2.
Lower than Ruth.
High national visibility.
High polarization.
The system considered Elena less dangerous because many people already viewed her through partisan or class narratives.
Naomi’s score was lower.
Claire’s lower still.
Ruth laughed.
“All your television interviews made you easier to ignore.”
Elena almost smiled.
The system had discovered what workers knew.
Fame could reduce trust.
Ruth’s ordinary life made her harder to dismiss locally.
Then Marcus warned them the transfer challenge had reached court.
Access could terminate any second.
Claire downloaded the audit logs.
Not the private household database.
They refused to copy vulnerable personal information unnecessarily.
Only operational metadata.
Who queried the models.
Who approved deployments.
The largest user account belonged to:
CV-14.
Catherine Vale?
Maybe.
Another:
EB-01.
Edward Bellmont.
Another:
MV-00.
Marcus Vale.
Marcus admitted it.
He built the early version.
No denial.
Then they found recent activity after Edward’s arrest and Catherine’s investigation.
Account:
RC-77.
Ruth Carter?
Ruth stared.
“No.”
Likely another planted identity.
They opened account registration.
Recovery contact:
Rosa Diaz.
Rosa.
The worker council leader who had removed Elena, Naomi, and Margaret from power.
The woman whose son’s cancer treatment Continuity once froze.
Nobody wanted to believe it.
That did not matter.
They had learned better.
Evidence.
Naomi called Rosa.
No answer.
The cooperative office said Rosa had left for a meeting two hours earlier.
Her destination was the Pennsylvania Department of Labor.
No meeting was scheduled.
Then Ruth received a text from Rosa.
I DIDN’T SELL YOU OUT.
A second followed.
THEY HAVE MY SON.
Everyone stood.
A third message.
No demand.
Only a photograph.
Rosa’s adult son sitting in a hospital room.
Conscious.
Uninjured.
An IV in his arm.
The hospital logo belonged to a Meridian client.
Then text appeared.
PHASE FOUR DOES NOT NEED KIDNAPPING.
IT ONLY NEEDS AUTHORIZATION.
Rosa’s son’s cancer medication had been placed under emergency utilization review.
Treatment stopped until an insurer approved it.
The same weapon.
Healthcare.
No physical captor.
No masked men.
A form.
A denial.
A clock.
Then the sender gave Rosa one instruction:
Use RC-77.
Authorize Pennsylvania deployment.
May you like
Rosa had not betrayed the workers.
Phase Four had turned her son’s treatment into a voting credential.
