Report 33 — The Morning After: Security Sector, Civil Service, Courts, and Administrative Continuity

Full report

On this page

Research question

How does a democratic transition keep society functioning while reforming institutions that may have participated in democratic erosion? Which personnel should remain, which should be removed, and how can machine intelligence help or harm this process?

Executive finding

A political uprising can succeed in removing leadership and still fail because the successor cannot govern. Book Three’s strongest operational insight is that food, water, hospitals, energy, communications, courts, records, pay, public safety, minority protection, and local administration do not pause when the square celebrates.

The transition must therefore solve two problems simultaneously:

continuity and accountability.

1. State capacity is not the same as regime loyalty

A common revolutionary mistake is to treat the entire state as the regime. Most public institutions contain a mix of political appointees, professionals, opportunists, passive employees, dissenters, and abusers.

Destroying the administrative state can turn a political victory into social collapse.

2. Civil service: presumptive continuity, targeted review

For ordinary civil servants, the default should normally be continuity unless evidence shows serious misconduct or a position itself was unlawfully politicized.

Useful principles:

• keep payroll functioning;
• prohibit political retaliation;
• review senior political appointments first;
• protect whistleblowers;
• document conflicts of interest;
• restore merit hiring;
• audit automated personnel systems for partisan criteria;
• allow appeals.

3. Courts present the hardest legitimacy problem

Backsliding governments often alter appointments, retirement rules, disciplinary systems, jurisdiction, or court composition. Undoing these changes can itself violate legal continuity if handled crudely.

Poland illustrates the dilemma: democratic recovery may require reversing politicized judicial reforms while avoiding the same arbitrary methods used to politicize the judiciary in the first place.

Prediction: judicial restoration will often take longer than electoral restoration.

4. Transitional judicial principles

A defensible restoration process can include:

• independent review of disputed appointments;
• transparent criteria;
• respect for final judgments except through lawful review;
• recusal/conflict rules;
• restoration of independent judicial councils where appropriate;
• discipline based on conduct, not ideology;
• temporary solutions to avoid case paralysis;
• constitutional adjudication of transition laws.

5. Election administration must be insulated early

An uprising or turnaround election is fragile if the next election is run by a partisan apparatus.

International IDEA’s studies of democratic backsliding show that election institutions can remain resilient even while other institutions deteriorate, but public confidence can be undermined by politicization and dispute-resolution failures.

Priority reforms include:

• clear appointment rules;
• staggered terms;
• professional staff;
• transparent procedures;
• auditable technology;
• public observation;
• independent dispute resolution;
• secure and public records.

6. Security sector: continuity without impunity

Police and military institutions cannot simply disappear. The question is whether they return to professional constitutional roles.

Reform priorities:

• civilian control;
• lawful command hierarchy;
• individualized vetting of senior and abusive personnel;
• professional promotion;
• clear rules for domestic deployment;
• independent investigation of unlawful force;
• inspector-general and legislative oversight;
• protection for lawful refusal of illegal orders.

7. Machine intelligence becomes part of the civil service

Government AI systems may decide or recommend:

• benefits;
• taxes;
• fraud investigations;
• immigration outcomes;
• hiring;
• procurement;
• risk prioritization;
• policing;
• health-resource allocation.

A transition that changes ministers but retains opaque models can preserve old discriminatory or political priorities.

Therefore the transition inventory must include models, datasets, system prompts, policies, vendors, and appeal processes.

8. The “model continuity” problem

Some automated systems may be essential to keep services running. Turning them off suddenly could cause harm. Leaving them untouched may perpetuate the old order.

A practical democratic approach is staged review:

1. identify high-impact systems;
2. preserve logs and documentation;
3. maintain essential service where immediate shutdown would harm the public;
4. suspend clearly unlawful uses;
5. conduct independent audits;
6. create human appeal;
7. replace or retrain systems under public rules.

9. Preventing old networks from hiding in the bureaucracy

Anti-democratic networks may persist through:

• contracting relationships;
• senior career posts;
• inspector positions;
• internal legal offices;
• procurement;
• security clearances;
• access to databases;
• party-linked public corporations.

The answer is not broad partisan dismissal. It is conflict-of-interest review, procurement transparency, access controls, rotation where necessary, and evidence-based vetting of sensitive posts.

10. The administrative “amnesty” that democracy requires

A successful transition needs ordinary people who served the previous state to believe they have a future under the new one. If every teacher, clerk, police officer, utility worker, and local administrator expects punishment, they have an incentive to sabotage reform.

The successor should clearly distinguish service under the old government from participation in abuse.

11. Strongest conclusion

The morning after is won by institutions that continue serving the public while becoming more accountable. Democratic transition is not the destruction of administration; it is the transfer of administration from party or ruler to law. Machine intelligence adds a new layer: the transition must audit not only officials but the automated systems through which officials govern.

Sources

1. Carnegie, Democratic Recovery After Significant Backsliding (2025): https://carnegieendowment.org/research/2025/04/democratic-recovery-after-significant-backsliding-emergent-lessons
2. International IDEA, Protecting Elections amid Democratic Backsliding in Poland (2025): https://www.idea.int/publications/catalogue/protecting-elections-amid-democratic-backsliding-poland
3. International IDEA, Parliamentary Rules in Democratic Backsliding and Resilient Design (2026): https://www.idea.int/publications/catalogue/parliamentary-rules-democratic-backsliding-and-resilient-design
4. DCAF, Post-war Security Sector Reform (2026): https://www.dcaf.ch/node/24385
5. OHCHR, Rule-of-Law Tools for Post-Conflict States: Vetting: https://www.ohchr.org/Documents/Publications/RuleoflawVettingen.pdf