Will AI Replace Archivists? The Surprising Truth About Digital Preservation

A faded letter from 1943. A set of architectural blueprints from a demolished building. A community newspaper that only ran for six months in 1978.

Deciding what gets preserved and what gets lost forever — that is what archivists actually do. And while AI is getting remarkably good at the mechanical side of this work, the judgment calls that define the profession remain deeply human. The data tells an interesting story about where exactly the line falls.

Where AI Is Already Making Inroads

[Fact] Cataloging and digitizing archival materials now has an automation rate of 62%. AI-powered optical character recognition (OCR) can transcribe handwritten documents with increasing accuracy. Computer vision systems can identify, tag, and categorize photographs by subject, era, and content. Automated metadata generation tools can process thousands of items in the time it once took to catalog a single collection box.

[Fact] Creating finding aids and metadata sits at 52% automation. This is the organizational backbone of archival work — the indexes, descriptions, and search tools that make collections usable. AI can generate draft finding aids, suggest classification codes, and identify relationships between documents across collections. The Library of Congress and National Archives have both piloted AI-assisted metadata generation with promising results. [Claim]

[Fact] Processing and preserving historical documents has an automation rate of 45%. Digital preservation workflows — format migration, checksum verification, storage management — are increasingly automated. But physical preservation decisions (how to handle a deteriorating daguerreotype, which adhesive to use on a crumbling map) remain hands-on.

The overall AI exposure for archivists stands at 41% in 2025, with a theoretical ceiling of 60%. [Estimate] By 2028, projections show exposure reaching 55% and automation risk climbing to 44%. This is firmly in the "medium" transformation category — significant change, but not wholesale replacement.

The Judgment Gap AI Cannot Close

Here is where it gets interesting. Archival work is fundamentally about appraisal — deciding what has enduring historical value. A box of corporate meeting minutes from the 1960s might seem unremarkable until an archivist recognizes it documents early environmental violations. A stack of personal letters might appear mundane until someone with historical context realizes they describe a neighborhood that no longer exists.

[Estimate] These appraisal decisions require cultural knowledge, historical context, community understanding, and ethical judgment that AI systems simply do not possess. An AI can tell you what is in a document. It cannot tell you why a future researcher in 2075 might need it.

The same applies to collection development — building relationships with donors, understanding institutional priorities, navigating the politics of what gets preserved and what does not. This is social work as much as information science.

[Fact] The BLS projects +4% growth for archivists through 2034. With roughly 7,400 workers in this field earning a median salary of about ,690, this is a small, specialized profession with modest but positive growth prospects. The digital explosion — more data being created than ever before — actually increases demand for professionals who can manage, curate, and preserve information over long time horizons.

This is an augment role. AI makes archivists faster and more capable, not redundant.

What Archivists Should Do Right Now

1. Master AI-assisted cataloging tools. Systems like ArchivesSpace, Archivematica, and emerging AI plugins are becoming standard. Archivists who can configure and quality-check these tools will be more productive and more employable.

1. Lean into the appraisal expertise. The more AI handles cataloging and metadata, the more your value concentrates in the irreplaceable judgment calls: what to collect, how to describe it, who it serves. Double down on this expertise.

1. Develop digital preservation skills. Understanding file formats, storage architectures, and long-term digital preservation standards (OAIS, BagIt, PREMIS) is increasingly essential. [Estimate] Institutions are struggling to preserve born-digital materials, and this need is growing exponentially.

1. Build community connections. Community archives, oral history projects, and participatory archiving are growing areas where human relationships and cultural competency matter enormously — and where AI has essentially no role.

1. Learn to work with AI transcription. Handwriting recognition, audio transcription, and translation tools are accelerating access to collections that were previously locked behind language or legibility barriers. Archivists who can leverage these tools to open up collections will create enormous value.

The future of archival work is not about choosing between humans and machines. It is about archivists who use AI tools to process the mechanical work faster so they can spend more time on the judgment, curation, and community engagement that only humans can do.

For detailed automation metrics, task-level breakdowns, and year-by-year projections, visit our Archivists occupation page. For comparison, see how AI affects related roles like librarians and museum curators.

Update History

• 2026-03-30: Initial publication with 2024-2028 data from Anthropic Labor Market Report.

Sources

• Anthropic, "The Anthropic Model of AI Labor Market Impact" (2026)
• Eloundou, T. et al., "GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models" (2023)
• U.S. Bureau of Labor Statistics, Occupational Outlook Handbook (2024-2034 Projections)

---

AI-assisted analysis. This article was generated with AI assistance and reviewed for accuracy. All statistics are sourced from peer-reviewed research and government data. For methodology details, visit our About page.