Skip to sources
Time Machine

Archivists

Scrub through 247years of this role's history, from when it first emerged, through every wave of technology that reshaped it, to the cited projections for where it's heading next.

2026drag to travel through time
180018251850187519001925195019752000now
Country
2026
Known today as Archivists (BLS SOC 25-4011) / Digital Archivist (born-digital specialization era)
Latest actual · 2024
9K
BLS OEWS May 2024, as reported via O*NET and the BLS National Employment Matrix. This is the present-day anchor used in projections. The 9,300 figure covers wage-and-salary archivists (8,400) plus self-employed archivists (approximately 1,000). The occupation remains small relative to its cultural footprint: fewer archivists are employed in the US than dental hygienists in any large city.
Latest actual · 2024
$61,570
BLS OEWS May 2024 median annual wage for archivists ($29.60/hr annualized). The wage distribution is skewed: federal government archivists (24% of employment) earn considerably above the median, while archivists at small nonprofits and local historical societies (a meaningful segment) earn well below it. The median has grown substantially in real terms since the 2000s as digital archivists with technical skills have commanded premiums and as federal pay scales have increased.
Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.
Tools of the era

The tools that defined the work

Select an era to see how it reshaped the work.

  • Physical custody: ledgers, acid-laden paper folders, wooden boxes

    The earliest American record keepers worked entirely by hand and physical organization. Records were stored in whatever came to hand: wooden boxes, leather portfolios, and later bound ledgers. The primary tool of the archivist was the hand-written inventory or calendar -- a list of documents in a collection with brief descriptions, produced by reading each item in turn. The labor was entirely physical and intellectual: no mechanical device assisted in organization, description, or preservation. The main threat to collections was fire, water, rats, and acid-containing paper that slowly destroyed itself. The keeper of records had no professional standards, no formal training, and no community of practice beyond correspondence with fellow historians.

    Ledger workPaper recordkeeping
  • Typewriter and card catalog (archival description enters the machine age)

    The typewriter, commercially available from the late 1870s, transformed the archival finding aid from a manuscript curiosity into a reproducible document. Card catalog systems, borrowed from librarianship, gave archives their first systematic subject and name access tools -- a typed card per folder or item, filed alphabetically, physically searchable. Repositories that adopted the typewriter could produce finding aids legible to any researcher; those that did not remained dependent on a single person's handwriting and memory. The technology did not speed the work dramatically but it standardized the output and made archives conceptually shareable across institutions for the first time.

    Work toolChanging equipment
  • Federal Records Acts and microfilm (preservation at scale)

    The founding of the National Archives in 1934 and the Federal Records Act of 1950 created the first formal institutional and legal framework for archivists. More transformatively, microfilm -- commercially mature by the late 1930s and widely adopted through the 1950s -- gave archivists their first tool for creating a preservation surrogate: a document could be filmed and the original could be managed (or destroyed under schedule) without losing access. The Eisenhower-era federal government microfilmed millions of pages of historical records. For the archivist, microfilm changed the job from mere physical custody toward active reformatting and preservation planning, the intellectual core of what the profession would become. The negative: microfilm introduced a new type of professional error -- poor filming procedures and vinegar-syndrome acetate bases created a preservation problem that archivists are still addressing.

    Effect on the work

    The Federal Records Act of 1950 and the supporting infrastructure of the National Archives system significantly grew the ranks of government-employed archivists through the 1950s. The national survey of archival agencies in the late 1950s counted roughly 1,300 repositories, each requiring at least one trained records professional.

    Work toolChanging equipment
  • MARC and mainframe databases (machine-readable bibliographic records)

    The Library of Congress's Machine-Readable Cataloging (MARC) format, developed in the late 1960s and widely adopted through the 1970s, gave archives their first chance to describe holdings in a computer-readable format. For archivists, the MARC-based RLIN and OCLC bibliographic systems (accessible via terminal by the mid-1970s) meant that finding a manuscript collection no longer required writing to each institution individually -- the records were increasingly discoverable online. The 1970s growth in social history -- oral history, labor archives, feminist archives, civil rights collections -- also drove a significant expansion of the archivist workforce as new repositories formed around newly valued historical communities. SAA membership grew substantially through this decade as college and university archives proliferated.

    Effect on the work

    The 1970s and early 1980s were a genuine growth period for archival employment: the rise of social history, new university archives, and state archival programs drove hiring that substantially expanded the occupation beyond its federal core.

    Mainframe processingComputerized records
  • EAD / internet-accessible finding aids (the World Wide Web transforms discovery)

    Encoded Archival Description (EAD), developed by the Library of Congress and SAA beginning in 1993 and released as Version 1.0 in 1998, was the first standard for encoding archival finding aids in a format designed for the World Wide Web. Before EAD, finding aids were paper documents that researchers had to request by mail or consult in person. After EAD, finding aids became internet-searchable, and the Archives Online platform and similar consortia made holdings visible to researchers globally for the first time. The archivist's job changed: the audience for their descriptive work exploded from in-person researchers to anyone with a browser, and the quality and completeness of description suddenly mattered in a new way. DACS (Describing Archives: A Content Standard), published in 2004, formalized the content standards that EAD encoding assumed.

    Effect on the work

    EAD did not directly expand archivist headcount, but it substantially raised the standard of what "good" archival description meant. The internet-visibility of holdings also revealed the scale of backlogs -- collections received but never described, physically safe but intellectually inaccessible -- that would eventually drive investment in digital-processing staff.

    Work toolChanging equipment
  • Born-digital processing: BitCurator, Archivematica, digital forensics

    By the mid-2000s, archives were receiving born-digital accessions -- hard drives, email exports, floppy disks, optical media -- alongside paper records, and had almost no tools to process them. BitCurator, a suite of digital forensics tools adapted from law enforcement by the University of North Carolina and Maryland (2011-2014), gave archivists their first professional-grade toolkit for acquiring, characterizing, and processing born-digital materials. Archivematica (2013) automated the preservation workflow end to end. These tools did not replace archivists; they created a new category of archivist, the "digital archivist," and a new body of technical knowledge required for the job. Institutions that could not hire digital archivists began to accumulate born-digital backlogs that, by the 2020s, ran to petabytes at major universities and federal agencies.

    Effect on the work

    Digital archiving skills became the most in-demand specialization in archivist job postings by the early 2010s, driving salary premiums and creating a supply gap the profession's graduate programs struggled to fill. By 2020, "Digital Archivist" was one of the fastest-growing job titles within the profession.

    Work toolChanging equipment
  • Transkribus AI handwritten text recognition (HTR)

    Transkribus, developed from the EU-funded tranScriptorium project (2013-2015) and the READ project (2016-2019), gave archivists their first practical AI tool for the most labor-intensive task in the manual-records world: transcribing historical handwriting. The platform grew from 2,200 registered users in 2015 to 45,000 in 2020, a twenty-fold increase in five years, with major institutions including the British Library, the National Archives of Estonia and Norway, and hundreds of university archives adopting it. For archivists, the impact was to convert transcription from a years-long project requiring armies of volunteers into a weeks-long supervised AI workflow: a custom HTR model trained on 75-150 pages of a specific hand can achieve 85-95% character accuracy on legible 19th-20th century cursive. The archivist's role shifted from transcriber to model trainer, quality controller, and exception-handler. Collections previously inaccessible due to transcription backlogs became discoverable for the first time.

    Effect on the work

    No archivist job loss has been documented as a result of Transkribus -- rather, the tool enabled archivists to tackle backlogs that would otherwise have required decades of unfunded labor. The profession's consensus is that HTR tools have expanded what one archivist can make discoverable rather than reducing headcount.

    Work toolChanging equipment
  • Generative AI tools: Whisper, LLM-assisted description, ArchivesSpace AI integrations

    The 2022-2025 generative AI wave arrived in the archives profession through several converging channels. OpenAI Whisper (September 2022) automated oral history and AV transcription with word error rates under 5% for clearly recorded English speech, turning multi-year AV backlogs into weeks of supervised review. ArchivesSpace, used by 500+ repositories, began integrating LLM-assisted drafting of scope notes and subject heading suggestions by 2024-25. ChatGPT and similar tools are being used by archivists to draft biographical histories, administrative histories, and grant narratives from structured inputs, though professional consensus holds that all LLM output requires rigorous verification before publication. The SAA formed an AI task force in 2023-24 to develop guidance. The net effect so far: augmentation, not displacement. The tasks that define the archivist's unique value -- appraisal authority, provenance judgment, access restriction decisions, physical authentication -- are precisely the tasks AI cannot reliably perform, and the profession's ethical framework makes human accountability for those decisions non-negotiable.

    AI audit toolsPattern detection
Projection cone · present → 2034

What credible sources project

Scrub the slider past now to anchor each scenario on the scrubber. The spread is the range of futures credible sources project for this role.

Employment outlook
Projected change in the number of people doing this work.
BLS Occupational Outlook Handbook -- Archivists, Curators, and Museum Workers (2024-25 edition)
2034
+6%
The BLS OOH 2024-25 edition projects 6% growth for the broader "archivists, curators, and museum workers" group (SOC 25-4011 plus 25-4012 and 25-4013 combined), with about 4,800 openings projected annually on average over the decade. The 6% group-level figure is slightly more optimistic than the archivists-specific 3.8% from the National Employment Matrix, reflecting stronger growth projected for curators and museum technicians within the group. The OOH attributes growth to increasing digital records, institutional digitization programs, and the expanding need to preserve electronic records in both public and private organizations. Reported here as a cross-check against the occupation-specific projection; the 3.8% archivists-only figure from the matrix is the more directly applicable one.
BLS National Employment Matrix 2024-34
2034
+3.8%
BLS Employment Projections -- industry-occupation matrix plus labor productivity assumptions. The 2024-34 cycle projects 3.8% employment growth for archivists (25-4011), from 9,300 to approximately 9,700 positions. This is classified as faster than average (all-occupations average is 3%). The BLS methodology models growing demand from public and private organizations needing to manage increasing volumes of electronic records, ongoing digitization mandates at federal and state archives, and the expanding challenge of born-digital records governance. The projection does not explicitly model the pace of AI adoption in archival workflows, which could either slightly dampen demand (if AI makes individual archivists dramatically more productive) or sustain it (as AI creates new complexity requiring human oversight).
AI task exposure
Share of the role’s tasks that researchers estimate AI can do. This is a measure of task exposure, not a forecast of jobs lost.
Frey and Osborne (2013) -- "The Future of Employment"
2033
76%
of tasks
Gaussian-process classifier on O*NET task features. Frey and Osborne placed archivists at 76% probability of computerization in their 2013 study -- a moderately high risk rating driven by the classification, organization, and retrieval tasks that dominated the archival job description in their O*NET input data. The 76% figure has aged poorly as a forecast: actual archival employment has grown since 2013, not declined, and the AI tools that arrived (HTR, automated metadata, Whisper) have augmented rather than replaced archivists. The bottleneck Frey and Osborne could not model was the appraisal task -- the irreversible, legally accountable expert judgment about which records have permanent value -- which is precisely the task AI cannot perform without human sign-off due to the catastrophic cost of errors. Reported here as the implied exposure ceiling under the F&O scenario, not as a realized or likely employment forecast.
Eloundou et al. -- "GPTs are GPTs" (2023)
2028
35%
of tasks
GPT-4 task-by-task LLM exposure labeling on O*NET tasks for archivists. Archivists score in the low-to-moderate range for LLM exposure -- the dominant high-value tasks (appraisal, access restriction decisions, physical authentication, provenance research requiring institutional context) require judgment and accountability that LLMs cannot reliably provide. The tasks most exposed to LLM augmentation are those archivists already use AI tools for: description drafting, transcription review, and metadata generation. Eloundou's framework measures LLM-specific exposure, not general automation: the archival tasks most threatened by AI tools are exactly those that AI tools have already transformed without reducing headcount. The 35% estimate reflects this partial exposure -- real but concentrated in tasks where augmentation (not displacement) is the operative dynamic.
Today, in this role

What's shifting in the work right now

The historical view above shows how this role has moved. This is the present-day detail: which AI tools are picking up which tasks, where the edge still is, and the natural directions this work can grow.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.

AI is sitting alongside you hereUse Transkribus AI handwritten text recognition (HTR) to transcribe historical manuscript collections at scale — training custom HTR models on an institution's specific handwriting styles, reviewing and correcting model output, and publishing corrected transcriptions that make collections discoverable for the first time without the years of volunteer labor previously required.

Use Transkribus AI handwritten text recognition (HTR) to transcribe historical manuscript collections at scale — training custom HTR models on an institution's specific handwriting styles, reviewing and correcting model output, and publishing corrected transcriptions that make collections discoverable for the first time without the years of volunteer labor previously required.[6],[7],[5]

Tools picking this up
Where your edge is

Transkribus HTR models trained on an institution's handwriting can achieve 85-95% character accuracy on 19th-20th century cursive and substantially higher on clearer hands — compressing what was once years of volunteer transcription effort into weeks of supervised AI processing. Your role shifts from transcriber to model trainer, quality controller, and exception-handler for difficult hands, damaged documents, and multilingual materials the model cannot handle. Learning to train custom Transkribus models and setting transcription quality thresholds is now a core archivist competency for any institution with significant manuscript holdings.

AI is sitting alongside you hereGenerate transcripts of oral history recordings and archival audio-visual materials using OpenAI Whisper or cloud-based AV transcription services — reviewing speaker-diarized output, correcting proper nouns and technical terminology, and time-stamping corrections for integration into AV access systems.

Generate transcripts of oral history recordings and archival audio-visual materials using OpenAI Whisper or cloud-based AV transcription services — reviewing speaker-diarized output, correcting proper nouns and technical terminology, and time-stamping corrections for integration into AV access systems.[8],[2]

Tools picking this up
Where your edge is

Whisper-large-v3 produces very high accuracy transcripts for clearly recorded audio in major languages (word error rates under 5% for standard American English), turning AV transcription from a months-long volunteer project into an hours-long review task. The accuracy gaps that require archivist correction cluster predictably: speaker attribution ("he said / she said"), proper nouns (names, organizations, place names), technical jargon, non-English passages, heavy accents, and overlapping speech. Build a post-processing workflow for Whisper output — spellcheck against a controlled vocabulary of names relevant to your collection — and you can process AV backlogs that were previously untouchable.

AI is sitting alongside you hereUse Adobe Acrobat AI and OCR tools to convert scanned documents into searchable PDFs — reviewing recognition accuracy for historical typefaces, correcting degraded-scan errors, and integrating corrected text into the repository's full-text search layer so researchers can discover collection contents that were previously invisible.

Use Adobe Acrobat AI and OCR tools to convert scanned documents into searchable PDFs — reviewing recognition accuracy for historical typefaces, correcting degraded-scan errors, and integrating corrected text into the repository's full-text search layer so researchers can discover collection contents that were previously invisible.[9],[2]

Where your edge is

Adobe Acrobat's AI-assisted OCR handles clean 20th-century typescript at very high accuracy rates and is increasingly effective on older typefaces, but degrades sharply on carbon copies, mimeograph duplicates, water-damaged documents, and non-Latin scripts. Your role is quality-sampling the OCR output at statistically meaningful intervals, building a remediation workflow for problem document types, and ensuring that OCR errors in finding aids and full-text layers don't systematically hide important materials from researchers. Train yourself on ALTO XML and HOCR formats so you can troubleshoot the text layer in your repository system.

Where this role is heading

Natural next steps for someone with your foundation: not exits, evolutions.

A direction you could grow

Education Administrators, Postsecondary

Education Administrators, Postsecondary (SOC 11-9033.00) — including university archivist director roles, head of special collections, and dean-of-libraries positions — represent the natural leadership escalation for experienced academic archivists. The pivot moves from processing collections to managing teams, budgets, and institutional strategy, with substantially stronger CRI because administrative authority and community trust are highly resistant to AI displacement. University archivists who step into director or dean roles are increasingly positioned as institutional AI policy advisors — deciding how AI-generated university records are retained, what HTR investments to make, and how to govern researcher access to sensitive digital collections. The barrier is administrative experience and a demonstrated leadership portfolio.

What you'd add
  • · Academic governance: faculty senate engagement, curriculum committee participation, accreditation self-study preparation
  • · Budget management and grant writing (NEH, IMLS, state humanities councils, private foundations)
  • · Personnel management and performance review in a unionized higher education context
  • · Strategic planning for library and archives programs: benchmarking against peer institutions, space planning, digitization roadmap development
  • · Institutional AI policy development: AI records governance, AI use policy for special collections, privacy framework for AI-assisted research tools
What it takesSome new skills to pick up
Share this year
Drops anyone you send it to straight into 2026.
Preview card
Part of Education · see all 51roles →
Different role?

See the same long-arc view for your own profession.

Browse the directory by industry, or search by title or SOC code. New roles ship every few weeks. Every profile cites every claim.

Browse all roles

The data behind this timeline

On record since1789
Latest tracked employment9,300 (US, 2024)
Latest median pay$61,570 (2024)
Outlook+3.8% by 2034 (BLS National Employment Matrix 2024-34)
View all 24 cited data points
YearUS employmentMedian annual paySource
193580n/aESTIMATE
19583,000n/aESTIMATE
20007,100n/aBLS-OEWS
20045,190$36,470BLS-OEWS
20055,410$37,420BLS-OEWS
20065,460$40,730BLS-OEWS
20075,420$43,110BLS-OEWS
20085,330$45,020BLS-OEWS
20094,900$46,470BLS-OEWS
20105,030$45,200BLS-OEWS
20115,460$46,750BLS-OEWS
20125,640$47,340BLS-OEWS
20135,560$49,110BLS-OEWS
20145,360$49,120BLS-OEWS
20155,460$50,250BLS-OEWS
20165,760$50,500BLS-OEWS
20176,080$51,760BLS-OEWS
20186,370$52,240BLS-OEWS
20196,560$53,950BLS-OEWS
20206,550$56,760BLS-OEWS
20216,120$60,050BLS-OEWS
20227,230$58,640BLS-OEWS
20237,150$59,910BLS-OEWS
20249,300$61,570BLS-OEWS
Embed this timeline on your site

Free for any site. Paste this where the timeline should appear; it stays interactive, every datapoint stays cited, and it sets no cookies on your page. How embedding works

<iframe src="https://futurehistory.earth/embed/25-4011"
  width="100%" height="430" style="border:0"
  title="Archivists, a Future History timeline"
  loading="lazy"></iframe>

See all roles in Education