AI Glossary / Data governance and quality / Intermediate
Link rot
Definition
Link rot is the loss of access to a resource through a previously useful web link, often because the resource has been removed, moved without a suitable redirect, or made unavailable at that address. A related problem is content drift: the address still works, but the content no longer matches what was originally cited. Both can undermine references, documentation, and digital evidence. A working HTTP response alone does not prove a citation still supports its claim.
What it means
A link identifies a location or resource through an address managed by someone. The publisher may reorganize a website, retire a service, change access conditions, or remove material. Other documents that rely on the old link do not automatically update. For a glossary or research article, this can leave readers unable to verify an explanation. Preserving useful citation details and reviewing important links reduces dependence on the continued behavior of one address.
How it works
For content you publish, design stable addresses and preserve meaningful redirects when pages move. For references, record enough descriptive information to identify the source independently, such as title, author or organization, and version. Where appropriate and permitted, retain a suitable archival reference or preserved copy. Check important citations periodically for both access and relevant content. Investigate failures before replacing links, and distinguish temporary outages or access restrictions from permanent removal. Keep historical source versions clear when newer material changes the claim.
Three practical examples
- Website maintenance: a glossary article's source link breaks after the publisher reorganizes its documentation.
- Project handover: an internal design document links to a retired workspace, leaving the next team unable to inspect the supporting decision.
- AI evidence: a retrieval system cites a live page whose content has changed since the answer was generated, making the original statement difficult to verify.
Think of it this way
Think of a footnote that gives a library shelf location but no title. If the library rearranges its shelves, the reference becomes hard to use. The analogy is limited because web pages can change their contents at the same address, redirects can lead elsewhere, and access can differ by time, region, or authorization.
Worked example
In an original review of 20 fictional citations, 16 return pages successfully, two return missing-page errors, and two require authentication. Among the 16 reachable pages, three no longer contain the information supporting the cited claim. Reporting 80% healthy references based only on successful responses would therefore be misleading. The reviewer records 13 currently reachable and relevant sources, three content mismatches, two missing pages, and two access-limited sources. The access-limited items are not automatically declared permanently lost. This example separates technical reachability, accessibility to a reader, and evidential relevance without treating any one status as a complete quality measure.
Illustrative case study
Fictional example: Aster Learning updates its AI glossary and finds that a source URL now redirects to a vendor's general marketing page. The page loads, but the original technical definition is gone. The editor locates an appropriate authoritative reference and records the change, while retaining the original citation details in the editorial history. This invented case shows why blindly accepting redirects can make a reference list appear healthy while weakening its support.
Common misconceptions and mistakes
Do not treat every temporary error as permanent link rot or assume every redirect is valid evidence. Do not replace a source with a similarly titled page without checking that it supports the same statement. An archived page may omit interactive content, downloads, or other dependencies, so inspect the capture. Also, a preserved historical source does not make its claims current. Keep source date, version, and access context distinct from the date on which the surrounding article was last edited.
Business use
Managing link rot protects the usefulness of educational content, technical documentation, and organizational memory. For an AI glossary, source links should support specific statements and be maintainable through later revisions. For retrieval systems, consider how answers identify the version or retrieval context of the evidence they used. Assign ownership for important reference checks and prioritize sources whose loss would materially affect trust or interpretation. Stable identifiers and archives help, but they still rely on stewardship and verification.
People also ask
Is a page that returns successfully necessarily a good citation?
No. Its content may have changed or no longer support the claim.
Is every login-required page a broken link?
No. It may be an access-limited resource, which should be described accurately.
Does an archive guarantee a complete copy?
No. Captures can fail or omit parts of a site, so the preserved result needs inspection.
Check your understanding
A source link redirects to a company's homepage. Should an automated checker mark the citation as fully verified?
Answer: No. It can record that the address resolves, but a relevance check is needed to establish whether the cited evidence remains available.
Related terms
- content drift
- reference rot
- persistent identifier
- permalink
- web archive
- digital preservation
- provenance
Learn next
audit five non-sensitive references in an old document. Record the original title, current destination, accessibility, and whether the relevant passage remains present. If you use an archive, inspect what was actually preserved and label its capture context. This makes citation maintenance a small evidence review rather than only a URL check.
More in Data governance and quality
- IT risk - IT risk is the possibility that uncertainty involving information technology will affect an organization's objectives.
- IT risk management - IT risk management is the ongoing process of identifying, assessing, responding to, and monitoring risks arising from an organization's use of information...
- Key management
- Mandatory access control - Mandatory access control, or MAC, is an access-control approach in which a system enforces security policy that ordinary users or resource owners cannot...
- Metadata - Metadata is information that describes, explains, organizes, or helps manage other information resources.
- Misuse detection - Misuse detection identifies activity by comparing it with descriptions of known unwanted or prohibited behavior.
Sources
Sources: Tim Berners-Lee, Cool URIs don't change, W3C, https://www.w3.org/Provider/Style/URI , relevant original-author excerpts consulted; Perma.cc documentation, https://perma.cc/docs , service explanation and capture-failure excerpt consulted; Perma.cc contingency plan, https://perma.cc/contingency-plan , relevant preservation-continuity excerpt consulted. These sources support address stewardship and preservation limitations. The citation audit and glossary case are original. No specific archive service's current pricing, universal capture completeness, or permanent availability is promised.
Previous: Linguistic value Next: Linkage constraints