Skip to content
STRATUSWORKFORCE SCAN

Curators

In short: today's best AI models could do about 12% of this job's working time (projected from METR's April 2026 measurement and Epoch's November 2026 index) if the work were set up for them, and about 48% by the end of 2028 if progress keeps its long-run pace. The people and hands-on work, about 48% of the time, stays with people.

Compare with another job

Administer collections, such as artwork, collectibles, historic items, or scientific specimens of museums or other institutions. May conduct instructional, research, or public service activities of institution.

12,150 people do this job in the US. $63,420 a year.

A typical 40-hour week

Not yet reviewed task by task
  • AI could take on5 h
  • A person checks AI's work0 h
  • Desk work AI can't do yet16 h
  • Stays with people19 h

Curators, now: of a 40-hour week, AI could take on 5 hours, 0 hours become checking its work, 16 hours are desk work it can't do yet, and 19 hours stay with people.

The largest pieces: Develop and maintain an institution's registration, cataloging, and basic record-keeping systems, using computer databases (about 3 h); Write and review grant proposals, journal articles, institutional reports, and publicity materials (about 1 h); Provide information from the institution's holdings to other curators and to the public (under an hour).

Estimates for a typical week of this job, if progress keeps its long-run pace, counting work the best models finish four times in five. Every task is below.

now
12%

of the working time. Now is projected from METR's April 2026 measurement and Epoch's November 2026 index.

Kinds of work
  • 52%
  • 7%
  • 41%
41%

of the work: Anthropic's measure of what people already bring to Claude at work (2025). It counts help with part of a task, so it can be higher than the share within reach.

Half the desk work within reach
October 2027

On the chosen pace and reliability. Estimate.

Share of the working time within reach, now to 2030

0%25%50%75%100%2027202820292030NowDesk work: 52% of the time
Chosen to Ceiling: the share
DateNow
NowEnd of 2030
Change assumptionsLong-run pace, four in five

What each choice means: , , , , . Every pace starts from .

At this date and pace, models finish tasks up to 2.4 work days long four times in five.

How far this has come

With GPT-4, measured in March 2023, 0% of this job's working time was within reach at . With Claude 3.7 Sonnet, measured in February 2025, 0%. With Claude Mythos Preview (early), measured in April 2026, 2%. Today's estimate: 13%.

0%5%10%15%20%2023202420252026GPT-4, March 2023: 0% within reachGPT-4 Turbo (Nov 2023), November 2023: 0% within reachGPT-4o, May 2024: 0% within reachClaude 3.5 Sonnet (June 2024), June 2024: 0% within reacho1-preview, September 2024: 0% within reachClaude 3.5 Sonnet (Oct 2024), October 2024: 0% within reacho1, December 2024: 0% within reachClaude 3.7 Sonnet, February 2025: 0% within reacho3, April 2025: 0% within reachGPT-5, August 2025: 0% within reachGemini 3 Pro, November 2025: 0% within reachClaude Opus 4.5, November 2025: 0% within reachGPT-5.2, December 2025: 0% within reachClaude Opus 4.6, February 2026: 1% within reachClaude Mythos Preview (early), April 2026: 2% within reachToday (estimate), November 2026: 13% within reach
Each step is a new best model measured by METR, four in five; the dashed end is today's estimate. Point at a dot for the model.

A backcast: today's task data with the best model METR had measured by each date. It shows how fast the models have moved, not what anyone forecast at the time.

Measured use over time

Of the Claude.ai conversations Anthropic sampled and tied to a work task, 0.085% went to this job's tasks in May 2026, the 180th largest of the 447 jobs that showed up at all; 2 of its 15 tasks showed up. Within each of Anthropic's methods, never across a change: 0.16% to 0.18% (released February and March 2025); 0.17% to 0.1% (August 2025 to February 2026); 0.079% to 0.085% (April to May 2026).

0%0.25%0.5%20252026released February 2025: 0.16%released March 2025: 0.18%the week of Aug 4, 2025: 0.17%the week of Nov 13, 2025: 0.098%the week of Feb 5, 2026: 0.1%April 2026: 0.079%May 2026: 0.085%
Each dot is one release's sample. A line joins releases made by the same method; a change of method starts a new line. Point at a dot for its window.
Each release
ConversationsShareRankTasks seen
Released February 20250.16%110 of 7087 of 15
Released March 20250.18%96 of 7036 of 15
The week of Aug 4, 20250.17%109 of 6556 of 15
The week of Nov 13, 20250.098%155 of 6726 of 15
The week of Feb 5, 20260.1%160 of 6777 of 15
April 20260.079%181 of 4342 of 15
May 20260.085%180 of 4472 of 15

Measured, not estimated: Anthropic's samples of Claude.ai conversations, each tied by a model to an O*NET task, from the Anthropic Economic Index. It says where Claude's use goes, not how much of this job's working time it covers, so it is not on the same scale as the reach above. The releases differ in their sample, their classifier and the smallest share they report (the June 2026 file rounds each task to a hundredth of a percent, so a smaller task reads as none), which is why a change of method starts a new line. At least 96% of each older release's task-tied use matched this site's tasks, and all of the newest. Claude.ai only, so use through other tools and other companies' models is not in it. Every job ranked.

Robots and the physical work

Anthropic rates 9% of this job's working time as physical work. For 4% of the working time, a robot has been shown doing most of the task somewhere: 4% in a structured facility. 0.07 of 3, the 547th highest of the 923 jobs.

  • Not physical 91%
  • No robot shown 5%
  • In a structured facility 4%

Anthropic's rating, not this site's estimate: Claude Opus 5, searching the web, rated each physical O*NET task by the kind of setting in which a robot it could cite (deployed, sold or demonstrated) does most of the task, weighted by its own estimate of each task's share of the time (What work can robots do?, September 30, 2026, data CC BY 4.0; every job ranked). A robot shown somewhere is not a robot in your building: cost, setup and safety decide that. The share within reach on this page still counts no body work. The robot behind each rated task is under the task below.

The tasks

Sorted by . 15 tasks from .

TaskKind
Plan and conduct special research projects in area of interest or expertise.Done more than yearly by most who do it. AI observed on this task.26%1.7 work weeks4.2 work days to 1.5 work months1%Half by Apr 2028
Plan and organize the acquisition, storage, and exhibition of collections and related materials, including the selection of exhibition themes and designs, and develop or install exhibit materials.Done more than monthly by most who do it. AI observed on this task.26%2.1 work weeks4.2 work days to 1.6 work monthsStays with people
Develop and maintain an institution's registration, cataloging, and basic record-keeping systems, using computer databases.Done more than monthly by most who do it. 20%6.7 h1.7 h to 4.2 work days40%Half by Jan 2027
Study, examine, and test acquisitions to authenticate their origin, composition, history, and to assess their current value.Done more than yearly by most who do it. AI observed on this task.Robots: no robot shown. Anthropic's rating.7%6.7 h2.5 h to 1 work weekStays with people
Write and review grant proposals, journal articles, institutional reports, and publicity materials.Done more than monthly by most who do it. AI observed on this task.5%6.7 h3.3 h to 4.2 work days61%Half by now
Train and supervise curatorial, fiscal, technical, research, and clerical staff, as well as volunteers or interns.Done more than weekly by most who do it. 4%2.5 h50 min to 6.7 hStays with people
Design, organize, or conduct tours, workshops, and instructional or educational sessions to acquaint individuals with an institution's facilities and materials.Done more than monthly by most who do it. 2%3.3 h1.7 h to 1.3 work daysStays with people
Attend meetings, conventions, and civic events to promote use of institution's services, to seek financing, and to maintain community alliances.Done more than yearly by most who do it. 2%3.3 h1.7 h to 1 work dayStays with people
Inspect premises to assess the need for repairs and to ensure that climate and pest control issues are addressed.Done more than yearly by most who do it. Robots: in a structured facility, Spot (with sensor payloads, Orbit fleet software) (Boston Dynamics) (deployed). Anthropic's rating.2%2.5 h50 min to 6.7 hStays with people
Negotiate and authorize purchase, sale, exchange, or loan of collections.Done more than yearly by most who do it. 2%6.7 h1.7 h to 3.3 work daysStays with people
Schedule events and organize details, including refreshment, entertainment, decorations, and the collection of any fees.Done more than yearly by most who do it.2%6.7 h1.7 h to 2.5 work daysStays with people
Provide information from the institution's holdings to other curators and to the public.Done more than monthly by most who do it. AI observed on this task.1%50 min13 min to 2.5 h99%Half by now

Within reach, per task: the share of the task's time whose instances are short enough for a model to finish at the chosen date, pace and reliability. Half by: the date half of that time comes within reach.

Jobs that share skills with this one

From O*NET's related occupations. Within reach now and by the end of 2028, long-run pace, four in five.

Questions about this job

How much of the work of Curators is within reach of AI now?
An estimated 13% of the working time is within reach of AI models now and 48% by the end of 2028, on METR's long-run pace at four in five (31% to 52% across the three paces the site shows). Within reach means a model could finish work that long at that reliability. It is not a forecast of jobs lost or of savings.
Which tasks of Curators come within reach first?
Weighted by the time each takes, the tasks with the most within reach now (projected from METR's April 2026 measurement and Epoch's November 2026 index) are: Develop and maintain an institution's registration, cataloging, and basic record-keeping systems, using computer databases (40% of the task's time now, 99% by the end of 2028); Write and review grant proposals, journal articles, institutional reports, and publicity materials (61% of the task's time now, 100% by the end of 2028); and Provide information from the institution's holdings to other curators and to the public (99% of the task's time now).
Which tasks of Curators stay with people?
Tasks that are people work or body work are not counted as within reach at any date. The largest: Plan and organize the acquisition, storage, and exhibition of collections and related materials, including the selection of exhibition themes and designs, and develop or install exhibit materials; Study, examine, and test acquisitions to authenticate their origin, composition, history, and to assess their current value; and Train and supervise curatorial, fiscal, technical, research, and clerical staff, as well as volunteers or interns. Together such tasks are about 48% of the working time.
Do people already use AI for the work of Curators?
In the Claude.ai conversations Anthropic sampled for May 2026 and tied to a work task, 0.085% went to this job's tasks, the 180th largest share of the 447 jobs that showed up; 2 of its 15 tasks appeared. Anthropic's labor market measure puts the share of the job's tasks seen in real Claude use at work at 41%. These are measures of use, not of how much of the work AI could finish, and they cover Claude only.
Can robots do the physical work of Curators?
Anthropic's robot exposure rating (September 2026) finds 9% of the working time is physical work, and that a robot has been shown doing most of the task, somewhere, for 4% of the working time, 4% in a structured facility. Its index is 0.07 of 3. One example: Spot (with sensor payloads, Orbit fleet software) for "inspect premises to assess the need for repairs and to ensure that climate and pest control issues are addressed". A robot shown somewhere is not one in every workplace; this site's own estimates still count no physical work as within reach.
How is this estimated?
From the 15 tasks O*NET lists for the occupation, each with an estimated length checked against GDPval and a share of the working time, each classed as desk, people or body work, and METR's measurements of how long a task the best AI models can finish, carried forward to the date from the latest, Claude Mythos Preview (early) in April 2026. Data checked November 10, 2026.

How the estimates are made

This job in your state or metro area
Embed these figures on your site

A small card with the share of working time within reach now and at the end of 2027, 2028 and 2030, the date and a link back. Free under CC BY 4.0; the card carries the credit.

See the card

Your own version of this job

The page above is the occupation. Your job is your title and what you actually do: give both, and they are read against these tasks in your browser, nothing sent anywhere. The headline is free: how much of your working time is within reach now and by the end of 2028, and how many of the tasks your description names.

This occupation has not been reviewed task by task yet, so the full personal report is not offered for it; 402 of 923 occupations are reviewed so far, largest first.