The Crisis of Custody
Unlike physical artifacts (shards, bones), digital artifacts are often entangled with living people. They contain "Personally Identifiable Information" (PII), copyrighted material, and sensitive secrets. To excavate without a filter is to be a grave robber.
The Three Primary Filters
The Custodial Filter operates as a decision tree:
1. The Privacy Filter (Protecting the Living)
Does the artifact contain PII? We distinguish between:
- Direct Identifiers: Names, emails, SSNs. (Must be REDACTED).
- Quasi-Identifiers: ZIP codes, birthdays, occupations which, when combined, create a unique fingerprint. (Must be assessed for context).
- Sensitive Attributes: Medical, political, or sexual data. (Requires EMBARGO or strict consent).
2. The Legality Filter (Respecting Ownership)
Do you have the right to copy this? Since most digital artifacts are owned, we rely on Fair Use (transformative research purpose) or document "Good Faith Effort" to locate owners of Orphan Works.
3. The Ethics Filter (Beyond Law)
Even if legal, is it right? We ask:
- Harm Assessment: Will this endanger anyone physically or reputationally?
- Community Consent: Did the community (especially marginalized ones) agree to be studied?
- The "Right to be Forgotten": Does the individual's right to delete their past override history's right to remember it?
Field Notes
The "Already Public" Fallacy: Just because something was posted on a public forum in 2005 does not mean the author consented to it being archived, searchable, and analyzed in 2025. Context matters. Privacy is contextual.
The Trolley Problem of Triage: You have 48 hours to save a dying platform. Do you save 10,000 random posts (comprehensive), or 100 at-risk posts from marginalized users (ethical priority)? The Custodial Filter forces you to choose explicitly.