Privacy-friendly analytics help us see which jobs people explore. They stay off unless you agree.Analytics off unless you agree. Details

Read the record
Back to the river
Skip to sources
Time Machine

Paralegals and legal assistants

Also called Paralegal, Legal assistant, Litigation paralegal, Corporate paralegal, Real estate paralegal

In brief

In May 2025, BLS counted 392,880 US paralegals and legal assistants, with median pay of $62,890 a year. Of the BLS counts since 2012, the May 2025 count is the highest. The earliest figure here is an estimate of 15,000 from 1975, not measured the same way as today's count. The last full technology era for paralegals and legal assistants: Casetext CoCounsel, 2023.

What the sources say

Official outlook

BLS projects little or no change in employment from 2025 to 2035 (-0.3%).

BLS Employment Projections 2025-35

AI exposure studies

  • 44% of work tasks in legal occupations, this job's group, are exposed to automation by AI. Not a forecast of job losses. Goldman Sachs, 2023
  • 45% of this job's tasks are exposed to large language models. Not a forecast of job losses. Eloundou and others, 2023
  • 0.94 estimated probability that the whole occupation could be computerised, over an unspecified number of years. Not a forecast of job losses. Frey and Osborne, 2013

Since the 2013 prediction

In 2013, Frey and Osborne put this job's probability of computerisation at 0.94. BLS counted 271,320 of these jobs in May 2013 and 392,880 in May 2025 (+45%, a change derived by Future History).

Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35. BLS publishes these counts as yearly survey snapshots and cautions against reading them as a precise trend. In May 2021 BLS changed how it estimates these counts (a model-based method using three years of survey data) and moved fully to the 2018 occupation codes, so part of a change across 2021 can come from the method rather than the work.

Last updated .

Cite this page
APA

Future History. (2026, October 11). Paralegals and legal assistants: jobs and pay since 1975 [Data set; data as of BLS OEWS May 2025; BLS Employment Projections 2025-35]. https://futurehistory.earth/timeline/23-2011

Chicago

Future History. "Paralegals and legal assistants: jobs and pay since 1975." Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35. Last modified October 11, 2026. https://futurehistory.earth/timeline/23-2011.

BibTeX

@misc{futurehistory_23_2011, author = {{Future History}}, title = {Paralegals and legal assistants: jobs and pay since 1975}, year = {2026}, month = oct, howpublished = {\url{https://futurehistory.earth/timeline/23-2011}}, note = {Last updated 2026-10-11. Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35}, }

The access date is today's date on your device.

drag to travel through time
19752000now
Country
2026
Known then as Legal Assistant (informal, pre-professional)
Known then as Legal Assistant / Paralegal (both in common use)
Known then as Paralegal and Legal Assistant
Known then as Paralegal and Legal Assistant (ABA definition updated, "legal assistant" removed)
Known today as Paralegal and Legal Assistant (ABA definition updated, "legal assistant" removed)
Projected to still be called Paralegal and Legal Assistant (ABA definition updated, "legal assistant" removed)
Data from the May 2025 survey
US Employment
393K
Est.

Stated in source: BLS OEWS

Official statistic
Report an error
Median Annual Wage
$62,890

≈ $61,278 in 2024 prices (derived by us, CPI-adjusted)

Official statistic
Report an error
Beat · 1967

The American Bar Association formally endorses the concept of the "legal assistant": a non-lawyer who can perform substantive legal work under attorney supervision. This is the founding policy moment of the paralegal profession.

Beat · 1968

ABA establishes its Standing Committee on Paralegals (originally "Committee on Legal Assistants"). The same year, formal paralegal education begins: eleven programs were operating by 1971.

Beat · 1975

National Association of Legal Assistants (NALA) incorporated April 19, 1975, in Minot, North Dakota. By March 1976, NALA had 1,200 members: the largest national paralegal organization at that time. The first Certified Legal Assistant (CLA) exam was administered November 1976 to 90 candidates; 42 passed.

Beat · 2012

Judge Andrew J. Peck issues the Da Silva Moore ruling (S.D.N.Y., Feb. 24, 2012): the first federal court order explicitly approving predictive coding / Technology-Assisted Review as a valid discovery method. The ruling opens the door for algorithms to replace paralegals in first-pass document review, the highest-volume paralegal work category.

Beat · 2023

Mata v. Avianca, Inc. (S.D.N.Y.): Attorney Steven Schwartz submitted a legal brief containing six entirely fictitious case citations generated by ChatGPT, including "Varghese v. China Southern Airlines" and "Estate of Durden v. KLM." Judge P. Kevin Castel sanctioned Schwartz and his firm $5,000 after they doubled down on the fake citations when challenged. The case became the canonical cautionary tale for unverified AI use in legal practice, cited by bar associations, malpractice carriers, and courts worldwide.

Beat · 2023

Thomson Reuters acquires Casetext for $650 million (deal signed June 26, closed August 17): the largest acquisition in legal technology history. Simultaneously, LexisNexis launches Lexis+ AI and Harvey reaches firmwide deployment at Allen & Overy. Within a single calendar year, every major legal research platform has released a generative AI product, and the incumbent incumbents have spent hundreds of millions to acquire AI capabilities.

Beat · 2024

BLS publishes the 2024-34 Occupational Outlook, projecting "little or no change" in paralegal employment over the next decade. For the first time in the profession's 56-year history, the BLS projection is flat rather than strong growth. The agency explicitly cites AI as the mechanism. Every prior BLS projection cycle (1980-2023) had projected double-digit decade growth.

How to read this number

OEWS is a point-in-time survey snapshot, not a continuous time series; BLS advises against using it for year-over-year trend comparison.

Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.

Official statistic: Copied by a script from a statistics agency table or a published study's table, not retyped by hand. BLS figures and the forecast and exposure tables are also re-checked against that table on every data run. Researched context: Researched from the cited source rather than copied from it by a script, so it is not machine-checked yet. Spot a mistake? Each figure has a “Report an error” link.

Cited forecasts · present → 2035

Will AI replace paralegals and legal assistants? What the research says

Each figure is stated by the named source; we publish no forecasts of our own. Where sources disagree, we show the disagreement rather than an average.

Each figure is stated by the named source; we publish no forecasts of our own. Where sources disagree, the spread is the honest signal.

What published forecasts say
Forecast change in the number of people doing this work, as published by each named source. Bars share one scale on every role, centred on zero and reaching the end at ±30%. Grey marks a small change, 3% or less either way.
BLS Employment Projections 2025-35
→ 2035
-0.3%
BLS National Employment Matrix, 2025-35 cycle, line for SOC 23-2011 (Paralegals and legal assistants): 404.9 thousand jobs in 2025, projected 403.8 thousand in 2035, a change of -1.1 thousand (-0.3%). Matrix employment counts all jobs in the occupation, including self-employed workers (2.1% of this occupation in 2025). The average for all occupations in this cycle is +3.5%.
Official statisticReport an error
How exposed is this work to AI (research measures)
Scores as published by each study. Each row says what it measures. These measure what AI could technically do, not forecasts of jobs lost.
Frey & Osborne (2013), "The Future of Employment"
2013 study · Probability of computerisation (2013 model)
94%
probability
What this measures

Published appendix value for SOC 2010 23-2011 Paralegals and Legal Assistants: probability of computerisation 0.94, rank 609 of 702 (rank 1 is least computerisable). Frey and Osborne estimated in 2013, with a Gaussian-process classifier on O*NET variables, how likely the whole occupation is to be technically automatable over "some unspecified number of years, perhaps a decade or two". It is a probability for the occupation, not a share of tasks and not a forecast of jobs lost.

Official statisticReport an error
Eloundou et al. (2023), "GPTs are GPTs"
2023 study · Share of tasks exposed to LLMs (β, human-rated)
45%
of tasks
What this measures

Published occupation-level exposure from the paper's occ_level.csv for O*NET-SOC 23-2011.00 Paralegals and Legal Assistants. Shown: human-rated β = 0.45, the share of the occupation's tasks an LLM could do in at least half the time at equal quality, counting tasks that need LLM-powered software at half weight (β = E1 + 0.5 × E2). Other published measures for this occupation: human-rated α (LLM alone) 0.05 and γ (all LLM-assisted) 0.85; GPT-4-rated β 0.525. This measures how much of the work an LLM could touch, not how many jobs disappear.

Official statisticReport an error
Goldman Sachs (Briggs & Kodnani, 2023), "The Potentially Large Effects of Artificial Intelligence on Economic Growth"
2023 study · Share of work tasks exposed to automation by AI, all legal occupations (US)
44%
of tasks
What this measures

Goldman Sachs Global Investment Research scored the task content of each US occupation for exposure to automation by generative AI and weighted by OEWS employment. Exhibit 5 (US panel) reports 44% for legal occupations, the occupational group this role belongs to; the report publishes no figure for individual occupations. It measures how much of the group's work AI could automate, not how many jobs disappear, and it is a 2023 estimate, not a forecast for a later year.

Official statisticReport an error
Today, in this role

What's changing day to day for paralegals and legal assistants

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

AI is sitting alongside you hereManage eDiscovery document review workflows in RelativityOne using aiR for Review: configure responsiveness and privilege review parameters, QA AI classifications on a statistically valid sample, and escalate borderline privilege calls to attorneys.

Manage eDiscovery document review workflows in RelativityOne using aiR for Review: configure responsiveness and privilege review parameters, QA AI classifications on a statistically valid sample, and escalate borderline privilege calls to attorneys.[12],[15]

Tools picking this up
Where your edge is

Earn the Relativity Certified Administrator credential. Shift focus from volume document review (which AI now handles) to designing the review protocol, sampling strategy, and QA workflow that makes the AI output defensible in court.

AI is sitting alongside you hereApply firm contract playbooks using AI redlining tools (Spellbook, Harvey) to review NDAs, vendor agreements, and real estate closing statements

Apply firm contract playbooks using AI redlining tools (Spellbook, Harvey) to review NDAs, vendor agreements, and real estate closing statements; escalate non-standard clauses to attorneys and maintain the firm's clause library.[13],[5],[16]

Tools picking this up
Where your edge is

Become the firm's expert in configuring and maintaining AI contract playbooks. Learn to write and refine custom review criteria in Spellbook Library so the AI reflects firm-specific standards, not generic templates.

AI is sitting alongside you hereReview and quality-check AI-generated first drafts of affidavits, legal correspondence, pleadings, and briefs produced by Harvey or CoCounsel before attorney review

Review and quality-check AI-generated first drafts of affidavits, legal correspondence, pleadings, and briefs produced by Harvey or CoCounsel before attorney review; flag hallucinated citations and factual inconsistencies.[9],[11],[17]

Where your edge is

Develop a structured QA checklist for AI-drafted documents: verify every citation against primary source, confirm party names and dates, flag any claim not traceable to the case file. Document errors to improve firm prompts.

Where paralegals and legal assistants go next

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

  • Compliance Officers
    US median pay
    $80,730 (May 2025)
    BLS outlook 2025-35
    +3.8%
  • Lawyers
    US median pay
    $159,670 (May 2025)
    BLS outlook 2025-35
    +4.7%

What each move takes

A direction you could grow

Compliance Officers

Paralegals already handle regulatory research, contract review, and policy interpretation: core compliance officer tasks. In-house compliance teams (fintech, healthcare, pharma) are actively recruiting candidates with legal research backgrounds and AI-tool fluency. The pivot typically requires domain knowledge in a specific regulatory framework (e.g., HIPAA, SOX, SEC rules) but not an additional degree.

What it takesMost of your skills carry over
Tools of the era

How has the work of paralegals and legal assistants changed since 1968?

The tools that defined the work, one line per era. Select an era to see how it reshaped the work.

  • Carbon paper, dictation machines, manual Shepardizing

    The first paralegals inherited the tools of the legal secretary: Dictaphone cylinders for capturing attorney instructions, carbon-paper copies for court filings, and the Shepard's Citations print index for checking whether a case was still good law. "Shepardizing", running a citation through the printed volume set, could consume an entire afternoon of a paralegal's working day. The job was essentially reading, copying, and reorganizing paper.

  • LEXIS (1973) and Westlaw (1975): electronic case research

    Mead Data Central launched the LEXIS full-text legal research service in 1973; West Publishing countered with Westlaw in April 1975. For the first time, a paralegal could search the entire corpus of US case law by keyword rather than by hand-turning index volumes. The technology was expensive and operated by dedicated terminal rooms in large firms through the 1980s; by the early 1990s it had moved to desktop PCs and displaced most manual citation-checking.

    Effect on the work

    Electronic research roughly halved the calendar time for comprehensive case research, but expanded the scope of research paralegals were expected to perform, so the net effect on hours worked was approximately neutral through the 1980s.

  • WordPerfect and dedicated word processors

    WordPerfect 4.0 (1984) became the dominant legal word processor of the 1980s: its dedicated legal-formatting macros for pleading paper, line numbering, and citation styles made it the default in law firms for a decade. The transition from IBM Selectric typewriters to word processors meant paralegals could produce, revise, and assemble documents without retyping from scratch, but also meant they absorbed the typing that had previously gone to the typing pool.

  • Email + the internet: correspondence and court filing

    Email supplanted interoffice memo routing and postal correspondence. Federal courts began piloting electronic filing (CM/ECF) in 1996; the system became mandatory in most federal districts by 2004. Paralegals absorbed the new role of e-filing coordinator: the person who understood the arcane technical requirements of each court's CM/ECF portal. Paper-based docketing and calendaring systems gave way to digital practice management software (ProLaw, Tabs3).

  • Concordance and Summation: early e-discovery document review

    Litigation support software (Concordance, Summation) gave paralegals their first systematic tools for organizing and searching large volumes of electronic documents. Before these platforms, document review in major litigation meant physically stamping and coding paper boxes; after them, it meant operating a database with full-text search. The volume of discoverable electronic data was also exploding, which more than offset any efficiency gain: e-discovery became a growth engine for paralegal headcount throughout the 2000s.

  • Legal Process Outsourcing (LPO): document review offshored to India

    Pangea3 (founded 2004, acquired by Thomson Reuters 2010) and dozens of Indian LPO firms offered document review. India's LPO sector grew between 2005 and 2009. US document-review paralegal positions that had been created by the e-discovery wave began migrating offshore or to contract staffing agencies, compressing wages at the bottom of the profession.

    Effect on the work

    The direct displacement of US paralegal document-review headcount is not precisely measured, but "first-pass review", the most labor-intensive paralegal task, was the primary product being offshored.

  • Relativity: cloud e-discovery platform (FRCP 2006 amendments)

    The 2006 FRCP amendments mandating disclosure of electronically stored information (ESI) created an overnight compliance requirement for every litigating company. Relativity, which began licensing in 2006, became the dominant e-discovery review platform. Paralegals in large firms became credentialed Relativity administrators: a specialization that commanded premium billing rates. The FRCP change also made e-discovery an unavoidable cost center, sustaining demand for review-layer paralegal headcount through the 2010s.

  • Predictive coding / Technology-Assisted Review (TAR)

    On February 24, 2012, US Magistrate Judge Andrew J. Peck issued his ruling in Da Silva Moore v. Publicis Groupe & MSL Group (No. 11 Civ. 1279, S.D.N.Y.), the first federal court opinion to explicitly approve the use of predictive coding (computer-assisted review) as a valid method for e-discovery. The ruling signaled that algorithms, not just human eyes, could satisfy discovery obligations. TAR adoption expanded through the 2010s; "first-pass" document review, a paralegal-dominated activity, began a slow migration to machine classifiers, displacing the most commoditized part of the paralegal workload.

    Effect on the work

    TAR did not eliminate paralegal e-discovery roles; it shifted them from volume-review coding toward quality-control, seed-set curation, and privilege review: higher-skill activities with less head-count leverage.

  • Spellbook: generative AI for contract drafting (GPT-3, then GPT-4)

    Rally (now Spellbook) launched in 2022 as the first AI tool built directly into Microsoft Word for contract drafting and review. Powered by GPT-3 at launch and GPT-4 thereafter, it could surface clause suggestions, flag missing provisions, and generate first-draft contract language: tasks that had been standard paralegal billable work in transactional practices. By 2025 it served thousands of in-house teams and law firms across many countries.

  • Harvey: GPT-4 legal AI deployed firmwide at Allen & Overy

    Harvey was founded in the summer of 2022 by Winston Weinberg (securities litigator) and Gabriel Pereyra (Google DeepMind / Meta). In February 2023, Allen & Overy, one of the world's largest law firms, announced it had deployed Harvey firmwide: thousands of lawyers had run tens of thousands of queries during the trial period. This was the first major Big Law endorsement of a generative AI tool for substantive legal work. Harvey reached an $11B valuation by March 2026.

    Effect on the work

    A&O's deployment was framed as augmenting lawyer productivity, not eliminating paralegal positions, but it demonstrated that LLMs could produce usable first drafts of legal memos, contract clauses, and research summaries at the speed paralegals bill by the hour.

  • Casetext CoCounsel: $650M Thomson Reuters acquisition

    Casetext launched CoCounsel in early 2023, a GPT-4-powered "AI legal assistant" that could complete document review, produce legal research memos, prepare deposition outlines, and analyze contracts in minutes. On June 26, 2023, Thomson Reuters announced it would acquire Casetext in cash: the largest M&A transaction in legal tech history and one of the ten largest AI acquisitions in the first half of 2023. The acquisition signaled that the incumbent legal research incumbents (Westlaw, LexisNexis) recognized that generative AI was an existential threat to their subscription model, not a feature to add later.

  • Lexis+ AI: LexisNexis generative AI platform (general availability October 2023)

    LexisNexis announced the beta of Lexis+ AI on May 4, 2023; it became generally available to US customers in October 2023. The platform combined conversational search, generative document drafting, and case summarization with Shepard's citation validation: explicitly positioning itself as "hallucination-free" by grounding outputs in the LexisNexis legal corpus. For paralegals whose primary billable work was legal research and memo drafting, Lexis+ AI automated the first-draft layer of both tasks.

Part of Legal · see all 6 roles →
Different role?

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

Browse the directory by industry, or search by title or SOC code.

Browse all roles

Paralegals and Legal Assistants employment and pay by year

On record since1968
Latest US employment392,880 (May 2025)Source: BLS OEWS May 2025Official statisticReport an error
Latest US median pay$62,890 (May 2025)Source: BLS OEWS May 2025Official statisticReport an error
Outlook-0.3% by 2035Source: BLS Employment Projections 2025-35Official statisticReport an error

In other countries

The latest figure each country's statistics agency publishes for this job. Each uses its own occupation definition, survey basis and currency, so the numbers are not directly comparable with each other or with the US figures.

  • Canada (NOC 42200)

    Employment, 2020
    31,205, Statistics Canada Census 2021 Official statistic Report an errorStatistics Canada Census of Population 2021: number of persons with employment income in NOC 42200 Paralegals and related occupations, Canada, reference year 2020. A recipients basis (anyone with employment income during the year), slightly broader than a point-in-time employed count. Statistics Canada Open Licence.
    Median pay, 2020
    C$62,000, Statistics Canada Census 2021 Official statistic Report an errorStatistics Canada Census of Population 2021 (Table 98-10-0586): median employment income for NOC 42200 Paralegals and related occupations, Canada, reference year 2020, full-year full-time workers (the basis comparable to the US OEWS full-time median). Statistics Canada Open Licence.

  • United Kingdom (UK SOC 3520)

    Employment, 2025
    41,000, ONS ASHE 2025 Official statistic Report an errorONS Annual Survey of Hours and Earnings (Table 14): employee jobs in SOC 3520 Legal associate professionals, United Kingdom, 2025 (ASHE "number of jobs", thousands x1000; excludes self-employment). Open Government Licence v3.0.
    Median pay, 2025
    £32,438, ONS ASHE 2025 Official statistic Report an errorONS Annual Survey of Hours and Earnings (Table 14.7a): median gross annual pay for all employee jobs in SOC 3520 Legal associate professionals, United Kingdom, 2025. Open Government Licence v3.0.

  • Australia (ANZSCO 5992)

    Employment, 2025
    22,400, ABS Labour Force Survey Official statistic Report an errorABS Labour Force Survey (Detailed), table EQ08: employed persons in ANZSCO 5992 Court and Legal Clerks, national (summed across sex and state/territory), annual average of 4 quarterly estimates.
    Median pay, 2025
    A$69,940, ABS Employee Earnings May 2025 Official statistic Report an errorAnnualised (x52) from A$1,345 median weekly earnings. ABS Survey of Employee Earnings and Hours, May 2025: median pay for full-time non-managerial employees paid at the adult rate, before tax. Surfaced via the Jobs and Skills Australia occupation profile.
View all 26 years of cited US data, each linked to its source
YearUS employmentAnnual pay (median unless marked)SourceReport
197515,000n/aEst., NALA History Researched contextReport
198698,000n/aEst., Wikipedia Researched contextReport
2000179,330$35,360BLS OEWS 2000 Official statisticReport
2003206,700$37,930BLS OEWS May 2003 Official statisticReport
2004210,020$39,130BLS OEWS May 2004 Official statisticReport
2005217,700$41,170BLS OEWS May 2005 Official statisticReport
2006229,430$43,040BLS OEWS May 2006 Official statisticReport
2007241,270$44,990BLS OEWS May 2007 Official statisticReport
2008253,040$46,120BLS OEWS May 2008 Official statisticReport
2009246,810$46,980BLS OEWS May 2009 Official statisticReport
2010247,940$46,680BLS OEWS May 2010 Official statisticReport
2011252,250$46,730BLS OEWS May 2011 Official statisticReport
2012267,030$46,990BLS OEWS May 2012 Official statisticReport
2013271,320$47,570BLS OEWS May 2013 Official statisticReport
2014272,580$48,350BLS OEWS May 2014 Official statisticReport
2015271,930$48,810BLS OEWS May 2015 Official statisticReport
2016277,310$49,500BLS OEWS May 2016 Official statisticReport
2017290,410$50,410BLS OEWS May 2017 Official statisticReport
2018309,940$50,940BLS OEWS May 2018 Official statisticReport
2019329,870$51,740BLS OEWS May 2019 Official statisticReport
2020332,720$52,920BLS OEWS May 2020 Official statisticReport
2021336,250$56,230BLS OEWS May 2021 Official statisticReport
2022345,240$59,200BLS OEWS May 2022 Official statisticReport
2023354,890$60,970BLS OEWS May 2023 Official statisticReport
2024367,220$61,010BLS OEWS May 2024 Official statisticReport
2025392,880$62,890BLS OEWS May 2025 Official statisticReport
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/23-2011"
  width="100%" height="430" style="border:0"
  title="Paralegals and Legal Assistants, a Future History timeline"
  loading="lazy"></iframe>
<p style="font-size:13px;margin:4px 0 0"><a href="https://futurehistory.earth/timeline/23-2011">Paralegals and Legal Assistants: jobs and pay over time</a>, from Future History</p>

See all roles in Legal