Foundations Series / Vol 01 Est. 2025

Chapter 9: The Custodial Filter — Ethics of Preservation

Opening: The Archivist's Dilemma

In 2018, a digital archivist received an anonymous hard drive in the mail. No return address, no note. Just a drive containing 50GB of data from a defunct online forum for survivors of domestic abuse.

The forum had shut down three years earlier when its volunteer admin burned out. No backup was ever released publicly. The community scattered, their stories lost. Until this drive appeared.

The archivist faced impossible questions:

Should she preserve this?

If she preserves it, who gets access?

Can she even contact the original posters?

What about the abusers mentioned in posts?

She sat with this drive for months, paralyzed. Every choice felt wrong.

This is the custodial burden—the weight of deciding what gets remembered and what gets forgotten, who gets privacy and who gets accountability, what serves history and what causes harm.

This chapter provides frameworks for navigating these impossible choices. Not easy answers (there are none), but ethical methodologies for thinking through preservation dilemmas systematically.

Part I: The Philosophy of Custodianship

What It Means to Be a Custodian

When you preserve a digital artifact, you become its custodian—responsible for its care, its interpretation, and its future.

This isn't neutral work. Preservation is an ethical act that:

1. Shapes Memory

2. Distributes Power

3. Affects Living People

4. Encodes Values

The Custodial Paradox

Custodianship involves fundamental tensions:

Preservation vs. Privacy

Accountability vs. Compassion

Comprehensiveness vs. Harm Reduction

Present Consent vs. Future Value

There are no formulas for resolving these tensions. Only frameworks for deliberation.

Part II: The Custodial Filter — Five Ethical Questions

The Custodial Filter (introduced in Chapter 5) is a systematic approach to preservation ethics. Before preserving any artifact, ask:

Question 1: Cultural Significance

Does this artifact matter?

This seems simple but is deeply contested. Who decides what "matters"?

The Traditional Canon Problem

Historically, archives preserved:

Marginalized communities were systematically erased:

Result: History is biased toward the powerful. Archives reflect and reinforce this.

Decolonizing Significance

New approach: Prioritize voices historically excluded:

Principle: If an artifact documents an underrepresented community or challenges dominant narratives, significance increases.

Community-Defined Significance

Best practice: Ask the community that created content whether it matters.

Example: Trans Archive Project (Hypothetical)

An archivist wants to preserve trans people's early YouTube videos (2006-2010). Before doing so, she:

  1. Contacts trans creators (where possible)
  2. Asks trans community members whether this is valuable
  3. Prioritizes what the community identifies as significant (not what she assumes)

Result: Preserves what trans people themselves think matters, not what outsiders imagine.

Challenge: Communities aren't monolithic. Trans people will disagree about what's important. Archivist must navigate plural perspectives.

Significance Over Time

What seems trivial today may be crucial tomorrow:

Principle: When uncertain, over-preserve. You can always restrict access later, but you can't un-lose deleted data.

Did the creator agree to preservation?

This is the hardest question because digital culture blurs public/private boundaries.

Explicit Consent:

Implied Consent:

Ambiguous:

No Consent / Violated Consent:

Sometimes, preserving without consent is justified:

1. Public Figures and Accountability

Politicians, CEOs, and public figures have reduced privacy expectations:

Example: Politician tweets racist statement, then deletes it. Archiving without consent is justified—accountability trumps desire to forget.

2. Historical Significance

Sometimes historical value outweighs individual privacy:

Guideline: The more significant, the more consent can be overridden—but never lightly.

3. Abandoned Content

If creator is unreachable (platform dead, email bounces, user vanished):

1. Private Content Leaked

Hacked emails, leaked DMs, stolen nudes—these should not be preserved, even if newsworthy:

Exception: If content reveals serious wrongdoing (corruption, abuse) and no other evidence exists—then restricted-access preservation with redactions might be justified. Case-by-case.

2. Content Explicitly Deleted

If a creator intentionally deleted something (not platform-deleted), that's a signal:

Exception: Public figures, accountability cases (as above)

3. Vulnerable Populations

Children, abuse survivors, people in crisis—their consent is especially important:

Guideline: Err on the side of respecting deletion/privacy for vulnerable people.

Ideal: Get explicit consent from everyone. In practice, this is often impossible:

Pragmatic approach:

  1. Prioritize consent when possible (contact creators if reachable)
  2. Assume implied consent for truly public content (but allow opt-out)
  3. Restrict access for ambiguous cases (preserve but don't make public)
  4. Don't preserve clear violations (leaked private content, explicit deletion by vulnerable people)

Question 3: Harm Assessment

Does preserving this artifact cause harm?

Some content, even if historically significant, causes ongoing harm by continuing to exist.

Types of Harm

1. Direct Physical Harm

Content that enables violence:

Principle: Do not preserve content that directly facilitates physical harm, even if "historically significant."

2. Psychological Harm

Content that retraumatizes:

Guideline: Preserve metadata (that it existed, summary of what it was) but not the content itself. Document without reproducing harm.

3. Reputational Harm

Old content that unfairly damages someone:

Trade-off: Reputational harm vs. accountability

4. Systemic Harm

Content that perpetuates oppression:

Complex Calculation:

Harm Mitigation Strategies

If you decide to preserve harmful content (for historical/research value), mitigate harm:

1. Restricted Access

2. Redactions

3. Contextualization

4. Opt-Out Systems

Example: Hate Forum Archive

A researcher wants to preserve a white supremacist forum (to study radicalization):

Harm Assessment:

Preservation Decision: Yes, but highly restricted

Harm Mitigation:

Question 4: Redundancy

Is someone else already preserving this?

If multiple institutions have copies, your effort might be better spent elsewhere.

Checking for Redundancy

Internet Archive's Wayback Machine:

Library of Congress Web Archive:

University Archives:

Community Archives:

Individual Creators:

When Redundancy Is Valuable

Even if something is archived elsewhere, you might still preserve if:

1. Different Preservation Methods

2. Institutional Fragility

3. Access Differences

4. Format/Quality Differences

When to Defer

If artifact is already well-preserved by stable institutions with good access:

Principle: Maximize coverage of endangered material; minimize duplication of secure material.

Question 5: Feasibility and Resource Allocation

Can you realistically preserve this, and is it the best use of resources?

Ethics isn't just about right vs. wrong—it's about triage under scarcity.

Resource Constraints

You have limited:

Ethical question: Given these limits, how do you allocate effort?

The Trolley Problem of Triage

Scenario: You have 48 hours before a platform shuts down. You can:

Option A: Preserve 10,000 posts from marginalized creators (high cultural value, small volume)

Option B: Preserve 1,000,000 posts representative sample of entire platform (lower per-post value, but comprehensive dataset)

Option C: Preserve 100 at-risk posts (doxxing targets, abuse survivors) that will cause harm if platform dies and they lose control of deletion

Which do you choose?

No right answer. Depends on:

Ethical Triage Principles

1. Prioritize the Endangered

2. Prioritize the Unrepresented

3. Prioritize Harm Reduction

4. Be Transparent About Trade-offs

5. Accept Imperfection

Grief is part of the work.

Part III: Case Studies in Custodial Ethics

Case Study 1: The Tumblr NSFW Purge (2018)

Background:

Custodial Dilemma:

Should archivists preserve purged content?

Arguments FOR:

Arguments AGAINST:

What Actually Happened:

Ethical Assessment:

Case Study 2: The January 6 Insurrection Videos (2021)

Background:

Custodial Dilemma:

Should archivists preserve deleted insurrection videos?

Arguments FOR:

Arguments AGAINST:

What Actually Happened:

Ethical Assessment:

BUT: Nuance Required

Case Study 3: The GeoCities Rescue (2009)

Background:

Custodial Dilemma:

Preserving without consent—ethical?

Arguments FOR:

Arguments AGAINST:

What Actually Happened:

Ethical Assessment:

Case Study 4: The Survivor Forum Hard Drive (Opening Scenario)

Background: Domestic abuse survivor forum (defunct 3 years), hard drive anonymously mailed to archivist

Custodial Dilemma:

What should archivist do?

Options:

Option A: Destroy

Option B: Preserve but Seal

Option C: Restricted Research Access

Option D: Try to Contact Posters

Ethical Analysis:

Competing Values:

Recommendation: Option B or C (Preserve but Restrict)

Reasoning:

  1. Destroy (A) loses valuable data (survivor narratives are historically underrepresented)
  2. Public access is wrong (clear privacy violation)
  3. Restricted access balances values:
    • Preserve for future (historical value)
    • Protect privacy (researchers must apply, IRB oversight)
    • Allow opt-out (if users emerge, they can request removal)
  4. Seal (B) or Research Access (C) depends on:
    • How identifiable are users? (If highly identifiable → seal)
    • How urgent is research need? (If active crisis → research access)

If choosing C (research access):

Lesson: When in doubt, preserve but restrict. You can always open access later (with community input), but you can't un-lose destroyed data.

Part IV: Building Institutional Ethics Frameworks

Creating a Preservation Ethics Policy

If you're building an archive or preservation organization, codify your ethical approach:

1. Mission and Values Statement

Example: "We preserve LGBTQ+ digital culture to ensure queer histories are not erased. We prioritize:

2. Preservation Criteria

What will you preserve?

How do you handle consent?

4. Access Tiers

Who can see what?

Tier 1: Public Access

Tier 2: Researcher Access

Tier 3: Community Access Only

Tier 4: Sealed

5. Takedown and Appeal Process

How do people request removal?

Automatic takedown for:

6. Ethical Review Board

Who makes hard decisions?

Example: Internet Archive's Approach (Simplified)

Mission: Universal access to knowledge

Consent: Respect robots.txt (if site owner says "don't crawl," they don't)

Access: Public by default

Takedown: Submit DMCA or personal information removal request, reviewed within days

Strengths:

Limitations:

Better model would add:

Part V: Personal Ethics for Individual Archivists

Your Own Custodial Ethics

If you're preserving as an individual (not an institution), you still need ethical frameworks:

1. Know Your Biases

Reflect:

Action:

Steps:

3. Document Your Decisions

Keep a log:

Reasons:

4. Seek Input

Don't decide alone:

5. Provide Escape Hatches

Build in ways for people to undo your preservation:

Part VI: When Preservation Is Wrong

Artifacts That Should Not Be Preserved

Some content should be actively destroyed, not preserved:

1. Child Sexual Abuse Material (CSAM)

Never preserve. Full stop.

If you encounter: Report to NCMEC (National Center for Missing & Exploited Children), delete immediately.

2. Non-Consensual Intimate Images (NCII / "Revenge Porn")

Never preserve.

If you encounter: Delete, report to platform or law enforcement if active

3. Doxxing That Endangers Lives

Do not preserve if:

Exception: If part of larger newsworthy event (Jan 6 insurrection), redact personal info but preserve rest

4. Terrorist Manifestos with Actionable Plans

Do not preserve if:

Exception: Preserve metadata (that it existed, summary) without full content. Work with law enforcement if actively dangerous.

5. Content Explicitly Illegal in Your Jurisdiction

Know the law:

If in doubt: Consult lawyer before preserving

Conclusion: The Weight of the Gavel

When you preserve, you wield power. You decide what future generations remember. You shape who gets voice and who gets silence. You determine what harms are perpetuated and what accountability is enabled.

This is not a burden to take lightly.

The Custodial Filter gives you structure for these decisions—five questions that force ethical deliberation:

  1. Does this matter? (Cultural significance)
  2. Do people consent? (Consent)
  3. Does preserving cause harm? (Harm assessment)
  4. Is someone else doing this? (Redundancy)
  5. Can I realistically do this? (Feasibility)

But structure isn't certainty. You will face dilemmas where all options feel wrong. You will make mistakes. You will carry the weight of artifacts you couldn't save and artifacts you shouldn't have saved.

This is the custodial burden.

Accept it. Feel it. Let it make you careful. But don't let it paralyze you.

Because the alternative—doing nothing—is also an ethical choice. And when platforms murder culture, doing nothing is complicity.

So preserve. But preserve ethically. Save what matters, respect those who don't want to be saved, restrict access when harm is possible, document your reasoning, and build escape hatches.

Be a custodian who wields power with humility, who makes hard choices transparently, and who accepts accountability for the consequences.

The gavel is heavy. Carry it anyway.

In the next chapter, we explore the Triage Workflow—the eight-step process for moving from discovery to preservation to access. Now that we understand the ethics, we'll learn the systematic methodology.

Discussion Questions

  1. The Survivor Forum: What would you have done with the domestic abuse survivor forum hard drive? Justify your choice using the Custodial Filter.

  2. Consent Boundaries: Where do you draw the line? At what point does historical significance override individual desire to be forgotten?

  3. Personal Bias: What are your own biases about what's "important" to preserve? How do those biases shape what gets remembered?

  4. Harm Weighing: How do you weigh potential future research value against present harm? Can you quantify that trade-off?

  5. Institutional vs. Individual: Should ethics be different for large institutions (Internet Archive) vs. individual archivists? Why or why not?

  6. The Custodial Burden: Have you ever made a preservation decision? How did it feel? What did you learn?

Exercise: Ethical Triage Simulation

Scenario: You're part of an archival collective. A platform announces shutdown in 72 hours. Your group has capacity to preserve 20% of the platform's content. Here are six types of content—rank them 1-6 for preservation priority, then justify your ranking:

Content Types:

A. Celebrity Accounts (1 million posts)

B. LGBTQ+ Youth Support Forum (50,000 posts)

C. Political Misinformation Archive (200,000 posts)

D. Fan Fiction Community (500,000 stories)

E. Indie Artist Portfolios (100,000 works)

F. Corporate Brand Accounts (500,000 posts)

Part 1: Ranking (300 words)

Part 2: Ethical Dilemmas (500 words)

For your top 2 choices, identify:

Part 3: Reflection (300 words)

Further Reading

On Archival Ethics

On Harm and Trauma-Informed Practice

On Digital Ethics

End of Chapter 9

Next: Chapter 10 — Triage Workflow: From Discovery to Preservation