COOKED
This residual category covers legal analysts, docket and compliance clerks, court program support, patent and trademark support staff, and similar unlicensed roles whose day is document review, docket tracking, records searches, form preparation, and summarizing case files — exactly the text-in/text-out work generative AI does at usable quality. Almost none of it requires a license or a signature, so no regulatory shield holds the work in human hands. The surviving slice is court- and client-facing coordination, chain-of-custody and filing accountability, and the judgment to spot when an AI-produced summary or citation is wrong before an attorney relies on it.
Roughly flat across the period, with year-to-year wobble.
Median pay $58,400 → $72,110 -1.2% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-1.2% 51,300 → 50,700 on the projections basis
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -1.2% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~4,700 openings a year on average, including replacing people who leave.
JurorCloserAdvocateBrief WriterCase BrieferLaw ExaminerEscrow CloserTariff ExpertContract ClerkEscrow OfficerLaw ResearcherTrial ExaminerPatent ExaminerPatent SearcherLegal TechnicianLegislative AideLife Care PlannerFriend of the CourtLegal AdministratorLegislative AssistantLegal Research Analyst
The BLS uses Legal Support Workers, All Other for work that doesn't fit any named occupation, so it covers roles that have little in common with each other. Two consequences worth knowing before you read anything below:
Holding it up: task resistance . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier A 7 rather than a 4 reflects the residue that still breaks automation — walking a physical exhibit through a clerk's window before a 4pm filing deadline, calling a county recorder whose index isn't online, reconciling a docket entry that PACER and the state e-filing system report differently — but the bulk of the day (privilege-log coding, deposition digests, Bluebook cite-checking, USPTO Office Action docketing, form 1040-style intake preparation) is already text-in/text-out.
Fully desk- and screen-based A 4 is the ceiling for desk work with an errand attached: courier runs to the courthouse, pulling boxed records from a file room, operating a scanner or Bates-stamping machine — physical, but climate-controlled, scheduled, and increasingly displaced by e-filing mandates.
No licence, no signature requirement A 3, not a 0, because docket and compliance clerks can be personally sanctioned for a missed statutory deadline or a certificate-of-service defect, but there is no bar admission, no notarial commission required in most postings, and the supervising attorney's signature — not yours — is what the court holds responsible under Rule 11.
Meaningful discretion A 7 sits at the bottom of real discretion: you decide whether a document is responsive or privileged on first pass, whether a conflicts hit needs escalation, whether a hallucinated citation in a draft gets flagged — consequential calls, but each one is reviewed by a licensed attorney before it leaves the building, and the deadline calendar itself is dictated by rule, not by your reading of ambiguity.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (7/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (3/20) is whether the law requires a licensed human to sign. Trust premium (6/20) is whether buyers specifically pay for a person. Judgment and accountability (7/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 16 of this occupation's 27 points (59%).
Embodiment (4/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 40/100 — EXPOSED.
USPTO practice: registered patent agents/practitioners already hold a signature-and-duty-of-disclosure obligation, and the USPTO's 2024 AI guidance requires a signer to personally review AI-assisted submissions. If patent/trademark support staff are pushed to obtain agent registration to sign, part of this SOC bucket acquires a genuine shield.
Genuine two-tier structure: as bulk summarization, docket entry and records search automate, what remains is adversarial verification — catching hallucinated citations, mismatched exhibit numbers, sealed-document handling, and jurisdiction-specific filing rejections. Task-mix shift alone raises the resistance of the residual role without any new rule, though it shrinks headcount.
Courts extending Rule 11-style AI certification duties down to support staff: several federal judges (e.g. standing orders in N.D. Tex., E.D. Tex. after Mata v. Avianca) already require a signed certification that AI-generated citations were human-verified. If bar rules or firm malpractice insurers require a named non-attorney verifier of record for each AI-drafted filing — as some insurers now condition cyber/malpractice riders on documented human review — the verification role becomes contractually mandatory rather than discretionary.
E-discovery chain-of-custody and sanctions exposure: if courts continue issuing spoliation sanctions where AI-assisted review missed responsive documents (FRCP 37(e) motions are rising), the custodian-of-record and privilege-log escalation calls become named, consequential decisions owned by support staff rather than clerical output.
The limit. No plausible route to a higher trust premium — clients buy attorney judgment, never the support layer, and nothing signals this changing. Embodiment is fixed near zero. Even with every lever, this is a much smaller occupation doing a narrower verification job; the levers raise the score of the survivors, not the headcount.
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 6,430 | $169,690 +135% |
| Chicago-Naperville-Elgin, IL-IN | 2,120 | $63,240 -12% |
| Los Angeles-Long Beach-Anaheim, CA | 2,020 | $80,330 +11% |
| New York-Newark-Jersey City, NY-NJ | 1,560 | $83,140 +15% |
| Denver-Aurora-Centennial, CO | 1,200 | $72,250 +0% |
| Atlanta-Sandy Springs-Roswell, GA | 1,060 | $62,740 -13% |
| Las Vegas-Henderson-North Las Vegas, NV | 900 | $60,820 -16% |
| San Francisco-Oakland-Fremont, CA | 840 | $102,420 +42% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 6,430 | $169,690 +135% |
| Charlottesville, VA | 40 | $139,240 +93% |
| Boulder, CO | 90 | $134,890 +87% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 27. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.