Opens in a new tab
needsahuman.

Will AI replace judicial law clerks?

A little.

Most of the work is research and drafting that AI can speed up, but the judge still needs a person to argue with. This job scores 67 out of 100 on (higher is safer). Today AI could do about 9% of the work by itself, people do 68% with AI’s help, and 23% still needs a person.

Updated 3 October 2026 23-1012 4131 2026-Q4
LegalJudicial Law Clerks23-1012 · 2026-Q4
9% AI does it68% AI helps23% needs a human
Your job's name, lit by the work that still needs a human.Needs a human 23%AI helps 68%AI does it 9%

AI does it: AI can do the task largely by itself. AI helps: a person still does it, faster with AI. Needs a human: AI can do little of it yet.

Why chambers work stays close to the judge

A judicial law clerk works for one judge, inside one chambers. The core of the job is reading, research, and writing for someone who has to decide a real case. A clerk reads the briefs and the record, finds the controlling authority, and writes a bench memo that says where the case actually turns. That memo is not a summary. It is a position the judge can question, in person, until the reasoning holds.

Two tasks explain the shape of this job better than any forecast. The first is legal research and citation checking, the slice that software has taken over fastest. The second is drafting opinions and orders under a judge’s direction, which is still a back-and-forth with the person whose name goes on the order. Question the draft, and the judge expects an answer, not a regenerated paragraph.

Scale matters here too. BLS counts about 13,290 judicial law clerks in the US, with median pay of $64,920 and projected employment change of 2.8% between 2025 and 2035 (BLS, 2025). Most clerkships are one- or two-year posts, so chambers can shrink quietly by hiring one clerk instead of two. That is the realistic pressure: fewer entry-level seats, not empty chambers. The same pattern shows up across early-career legal roles on our entry-level hiring tracker.

What AI drafts, what it assists with, and what the judge still needs

Start with the tasks software can run with little supervision. The share of task time in that group: 9%. It covers searching statutes and case law for on-point authority, and checking quotations and citations against the sources they came from. Both are pattern work with a right answer, and both used to eat a clerk’s first week on a file.

Next, the assisted group, where a clerk stays in charge and the tool shortens the task. Its share of task time: 68%. Summarizing a long record or a thick set of briefs sits here, and so does producing a first draft of a memo or proposed opinion. The draft is a starting point. The clerk still verifies every cite, because the court holds a person responsible for what the document says. Our coverage method page explains how that task time is measured.

Then the work that stays with a person. Its share: 23%. Discussing a case with the judge and recommending an outcome belongs there, along with confidential deliberation and the courtroom and conference duties that come with the clerkship. These tasks are not hard because they are technical. They are hard because they require a trusted person who can be questioned, corrected, and held accountable inside a closed chambers.

What has actually been tested

No study in our evidence set has put AI drafting against judicial law clerk work and measured the result. The evidence grade for the quality comparison on this page is D, and a D grade means no parity number is published for this job. We do not estimate one from nearby roles.

What would settle it is specific. A blind review in which sitting judges rate clerk-written bench memos against AI-written drafts on the same files. Measured error rates for automated citation checking across real dockets, not sample sets. Published results from court pilots that name the tool, the task, and the verification step. Until something like that exists, the honest statement is that the research and cite-checking slice is clearly assisted, and the advisory slice is untested. How grades are assigned is set out on the quality parity page, and the full approach sits at our methodology.

When the balance could shift

Most likely between 2036 and 2049 (8 in 10 of our scenarios). The replacement-year method page explains what that window does and does not mean.

Two things could pull the date earlier. Research and drafting tools are being built directly into court case-management systems, so the work moves to the tool without anyone choosing a new product. And judiciary budgets are tight, which makes a smaller clerk pool an easy saving when software covers the first draft. Nothing physical stands in the way either: the robotics requirement for this job is listed as none needed, so there is no hardware cost to clear.

Two things hold it back. Courts place the duty to verify on a named person, so every citation an AI produces still needs human sign-off, which removes much of the time saved. And a clerkship is partly a training post. Judges hire clerks to build the next generation of advocates and judges, a purpose no software serves. You can see the same split in the blockers and costs listed above, and compare it with a neighboring role on our job comparison tool.

How to stay needed in chambers

Lean into the tasks in the needs-a-human group. Be the clerk who can defend a recommendation out loud when the judge pushes back. Take on the confidential, sensitive files where judgment about parties and credibility matters more than retrieval. Own the courtroom and conference work, where you see how a case behaves in front of real people rather than on paper.

Two skills carry the most weight. First, verification discipline: a documented habit of checking every authority, quotation, and record cite, including anything a tool produced. Second, clear written reasoning under time pressure, the kind that survives an appeal and reads well to a judge who has twenty other files open.

What to do: keep a short record of the judgment calls you made on each file, not just the research you ran, so your value is visible when chambers budgets are set.

If you are weighing where a clerkship leads, the closest work sits with Lawyers, Judges, Magistrate Judges, and Magistrates, and Administrative Law Judges, Adjudicators, and Hearing Officers. Support roles such as Paralegals and Legal Assistants share much of the same research and drafting exposure. For the wider picture, see the lawyers and judges family, the government sector page, or look this role up beside others in the full job rankings.

Frequently asked questions

Is a judicial law clerk a lawyer?

Usually, yes. Most federal and state judicial law clerks are law school graduates, and many are admitted to a bar. The clerkship is a term appointment, often one or two years, spent researching and drafting for a single judge. Some courts also employ career clerks who stay for years. Court clerks who handle filings and dockets are a separate occupation with different duties.

Will AI take over judges' jobs?

Judging carries legal authority that software cannot hold. A judge is accountable for the decision, must hear parties, and can be appealed. AI can draft text and surface authority, but it cannot take responsibility for an order. Judges also have their own page on this site, with their own task split and evidence, so you can see which parts of that job are assisted and which are not.

Can a law clerk become a judge?

Many judges clerked early in their careers, but a clerkship is not a direct path to the bench. Judges are elected or appointed, usually after years of practice as a lawyer, prosecutor, or in another legal role. A clerkship helps because it teaches how courts reason and decide, which selection committees and appointing authorities value.

Which legal tasks has AI already absorbed?

Searching case law and statutes, pulling on-point authority, drafting first-pass summaries of long records, and checking citations and quotations against sources. The time saving is real, but a person still verifies the output because courts hold a named individual responsible for what a filing says. The task list above shows which chambers tasks sit in the assisted group and which do not.

Will AI eventually replace paralegals?

Paralegal work overlaps with clerk work on research, document review, and cite-checking, so it carries similar exposure. The pressure shows up first in hiring: fewer junior seats rather than existing staff leaving. Client contact, case organization, deadline management, and courtroom support remain human tasks. The paralegal page on this site sets out that task split and the evidence behind it.

How many judicial law clerks work in the US?

BLS counts about 13,290 judicial law clerks, with median pay of $64,920 and projected employment change of 2.8% from 2025 to 2035 (BLS, 2025). It is a small occupation tied to the number of judgeships and court budgets, so changes tend to come through hiring decisions in individual chambers rather than broad industry shifts.

Each ridge is a slice of the job's task time.Needs a human 23%AI helps 68%AI does it 9%
The job’s mark

No two jobs leave the same print

Every job gets its own fingerprint, drawn from its code. The amber ridges are the share of task time that still needs a person. Below them, the same ridges are written out in ones and zeros: slate for the work AI helps with, white for the work AI can do.

Judicial Law Clerks, O*NET-SOC 23-1012. 23% of the job’s task time still needs a human, so 23 of every 100 ridges are amber; slate is what AI helps with, white what AI can do.

What AI can and cannot do

The tasks that make up the job, from , and where AI stands on each today: , (a person does it, with AI speeding it up) or . 23% of the still needs a human.

Each block is one task; its height is its share of working time.Needs a human 23%AI helps 68%AI does it 9%
The job's task list: the parts AI can do are blacked out.Needs a human 23%AI helps 68%AI does it 9%
Prepare briefs, legal memoranda, or statements of issues involved in cases, including appropriate suggestions or recommendations.AI helps
Research laws, court decisions, documents, opinions, briefs, or other information related to cases before the court.AI helps
Draft or proofread judicial opinions, decisions, or citations.AI helps
Confer with judges concerning legal questions, construction of documents, or granting of orders.Needs a human
Review complaints, petitions, motions, or pleadings that have been filed to determine issues involved or basis for relief.AI does it
Keep abreast of changes in the law and inform judges when cases are affected by such changes.AI helps
Attend court sessions to hear oral arguments or record necessary case information.Needs a human
Review dockets of pending litigation to ensure adequate progress.AI helps
Communicate with counsel regarding case management or procedural requirements.AI helps
Respond to questions from judicial officers or court staff on general legal issues.AI helps
Enter information into computerized court calendar, filing, or case management systems.AI helps
Verify that all files, complaints, or other papers are available and in the proper order.AI helps
Coordinate judges' meeting and appointment schedules.AI helps
Participate in conferences or discussions between trial attorneys and judges.Needs a human
Prepare periodic reports on court proceedings, as required.AI helps
Supervise law students, volunteers, or other personnel assigned to the court.Needs a human
Maintain judges' law libraries by assembling or updating appropriate documents.Needs a human
Perform courtroom duties, including calling calendars, administering oaths, and swearing in jury panels and witnesses.Needs a human

Is it better than a person? The evidence

No direct test against people in this job yet. Every study is , and vendor studies are labelled as such.

When could it be replaced?

When AI could largely do this job: 2036–2049

Most likely between 2036 and 2049 (8 in 10 of our scenarios). A range from our of how fast AI improves, how fast employers take it up and what holds it back, not a forecast that the job ends. “” has a strict meaning here. Today’s answer is at the top of the page; this is how it could change.

The sand is the human working years left, measured in the same 40-year glass for every job, so a safe trade starts nearly full and an exposed job with a thin layer.

The sand is the human working years left, in the same 40-year glass for every job.Years still needing a humanYears run out

How this job could shift, year by year

Where the job could sit on our scale each year to 2060, across the ten behind its .

Today
Will AI replace this job?
A little.
By 2045
90%
of our scenarios have AI largely doing this job by 2045 (Largely.)
0% still have it mostly needing a person (A little. or Nah.)
By 2060
100%
of our scenarios have AI largely doing this job by 2060 (Largely.)
0% still have it mostly needing a person (A little. or Nah.)

We run this job as ten scenarios spread across its replacement range. In each, the score moves towards the bottom band (Largely: AI could largely do the job) by the year that scenario reaches it, slowly at first and faster later, as adoption usually goes. Each bar splits the ten by the band they put the job in. The model stops at 2060. How the timeline works

Share of this job's scenarios in each verdict band, today to 20600%25%50%75%100%2026: 100.0% of scenarios: AI could do a little of this job (A little.)100%Today2030: 40.0% of scenarios: AI could do a little of this job (A little.)40%2030: 60.0% of scenarios: AI could partly do this job (Partly.)60%20302035: 40.0% of scenarios: AI could partly do this job (Partly.)40%2035: 40.0% of scenarios: AI could mostly do this job (Mostly.)40%2035: 20.0% of scenarios: AI could largely do this job (Largely.)20%20352040: 40.0% of scenarios: AI could mostly do this job (Mostly.)40%2040: 60.0% of scenarios: AI could largely do this job (Largely.)60%20402045: 10.0% of scenarios: AI could mostly do this job (Mostly.)10%2045: 90.0% of scenarios: AI could largely do this job (Largely.)90%20452050: 100.0% of scenarios: AI could largely do this job (Largely.)100%20502055: 100.0% of scenarios: AI could largely do this job (Largely.)100%20552060: 100.0% of scenarios: AI could largely do this job (Largely.)100%2060
Will AI replace the job?Largely.Mostly.Partly.A little.Nah.
Share of this job's scenarios in each band, year by year. Updated with every release.
Show the data
YearLargelyMostlyPartlyA littleNah
Today (2026)0.0%0.0%0.0%100.0%0.0%
20300.0%0.0%60.0%40.0%0.0%
203520.0%40.0%40.0%0.0%0.0%
204060.0%40.0%0.0%0.0%0.0%
204590.0%10.0%0.0%0.0%0.0%
2050100.0%0.0%0.0%0.0%0.0%
2055100.0%0.0%0.0%0.0%0.0%
2060100.0%0.0%0.0%0.0%0.0%

What’s stopping AI taking over?

The things that keep this work with people, strongest first. Each is scored 0 to 100 from work context, licensing and the evidence we have.

LicensingUsual entry requirement (BLS): doctoral or professional degree.
LiabilityMistakes are rated 2.7 out of 5 for consequence and decisions 4.6 out of 5 for impact; someone has to answer for them.
Evidence gapNo study yet compares AI with people doing this job, so employers have no proof it is good enough.
Clients want a personFace-to-face contact is rated 4.6 and physical closeness 2.5 out of 5; caring for or serving people is 2.6 out of 5 in importance.
RegulationWorkers rate responsibility for others' health and safety 1.3 out of 5; the sector has its own rules on who may do the work.
Physical work2% of the task time is physical; robots have been shown on 100% of that time.

What would it cost to hand the work to AI?

The share of the year AI could handle (745 of 2,080 hours a year), priced two ways. Both are ranges, not quotes.

AI model usage, a year
$70–$7,450
A person’s wage for the same hours
$16,530–$40,500

AI cost covers model usage only: no integration, licences, oversight or the human time still needed to review the work. Human cost is the wage for the same hours, without benefits or overheads. As of 2026-10.

Robots and humanoids

AI software can only take the work at a screen. The rest needs a robot that can do it.

2%
of the task time is physical work
None needed
the kind of robot the physical work would need
Little of this job is physical, so robotics is not what holds AI back.

Source: Anthropic Economic Index, 'What work can robots do?' (30 September 2026); O*NET 31.0 task weights.

Which AI skills does this job lean on?

The job’s task time split by what an AI model would need to be good at, and where models stand today.

Each star is a task, grouped by the AI skill it leans on.Needs a human 23%AI helps 68%AI does it 9%
Writing · 27.1% of time
Strong
Drafts, edits and translates most routine documents at professional quality.
Analysis · 29.3% of time
Strong
Reliable on structured data and rules; uneven on judgement calls with thin information.
Coding · 0% of time
Strong
Agents complete many routine software tasks end to end; larger systems still need people.
Vision and design · 5.2% of time
Good
Reads documents, images and layouts well; specialist imaging needs dedicated, approved tools.
Speech · 17% of time
Good
Voice agents handle routine calls and live interpreting; complex or sensitive calls still go to people.
Planning and agents · 17.3% of time
Emerging
Multi-step agents work in narrow, well-tooled workflows; open-ended coordination is unreliable.
Physical manipulation · 1.8% of time
Early
Robots handle structured, repetitive handling; general dexterity outside fixed settings is not commercial.
Care and persuasion · 2.3% of time
Limited
Can script, coach and advise; trust, presence and accountability still need a person.
Amber matter holds its orbit, slate circles the inner disc, white falls in.Needs a human 23%AI helps 68%AI does it 9%
How exposed is it?

Still needs a human: 67/100↑ safer

The amber matter on the outside holds its orbit: that is the task time that still needs a person. The slate matter circling the inner disc is the work AI helps with. The white matter is the share AI can do; it spirals in and is gone over the edge.

The matter follows the job’s task split: 23% needs a human, 68% AI helps, 9% AI does it. Still needs a human: 67/100 ↑ safer. Will AI replace them? A little.

What the AIs say

We asked four AI assistants the same question: will AI replace this job in the next 10 years? One word (Yes, Partly or No), then one sentence. Our own answer, from the data: Still needs a human: 67/100 ↑ safer. Will AI replace them? A little.

ChatGPTPartly

AI will likely automate research, drafting, and document-review tasks, but judges will still need human clerks for judgment, ethics, discretion, and nuanced legal reasoning.

gpt-5.5-2026-04-23 · asked 2026-10-03
ClaudeNo

While AI will increasingly handle research and drafting support tasks, the nuanced legal judgment, ethical reasoning, and trust required in judicial decision-making will keep human law clerks essential for the foreseeable future.

claude-sonnet-5 · asked 2026-10-03
GeminiPartly

While AI will automate routine tasks like legal research and draft preparation, human clerks will remain essential for nuanced legal reasoning, ethical judgment, and close collaboration with judges.

gemini-3.8-flash · asked 2026-10-03
PerplexityPartly

AI will likely eliminate many routine clerk tasks and reduce some positions, but human clerks will remain necessary for judgment, verification, confidentiality, and nuanced legal analysis.

sonar · asked 2026-10-03

Assistants answer from what they learned in training (Perplexity also searches the web), so they can be confidently wrong, and the same question can get a different answer tomorrow. Our score is built from task data and graded evidence. Answers collected through DataForSEO.

Cite this page

NeedsAHuman.com (2026). Will AI replace Judicial Law Clerks? A little. Still needs a human: 67/100, higher is safer; release 2026-Q4. https://needsahuman.com/jobs/judicial-law-clerks/ (accessed 4 October 2026).

Scores change with each , so cite the release. The data is open under : credit NeedsAHuman.com with a link. Open data · Press

Put the badge on your site

The badge updates itself with each release and links back to this page.

Sources

  • Tasks and work context: 31.0, ().
  • Jobs, pay and projections: US , and 2025–35.
  • How AI is used today: ; Microsoft Research, .
  • What AI can do: our task ratings ( r1) and the quality evidence register.
  • UK names and employment: coding index and .

How each score is built: methodology. Every figure on this page: open data. Release 2026-Q4.