# openai-work-at-the-frontier-task-crossover-2026-07-27

## Veille

Post and report from **OpenAI Economic Research** published on **July 27, 2026**, the first installment in the **Work at the Frontier** series, based on an analysis of **more than 800,000 messages from U.S. ChatGPT users**. **Coined concept**: ***task crossover*** — *« work historically associated with one occupation appearing in the AI use of people in another »*. **The headline figure is actually two figures, and that's the point most coverage loses**: **16.8% of work-related messages** concern tasks associated with a different occupation, and **43.5% of occupation-specific messages**. The funnel explains the gap: **61.5% of usage is generic** (writing, summarizing, planning — too widely shared to count as evidence of crossover) and is excluded; of the **remaining 38.5%**, **43.5% fall outside the occupation** and 56.5% are *« inside **or near** »* — the upper bound is thus calculated on a reduced base, the lower bound on the entire work-related usage. **By occupation** (share of occupation-specific messages pointing to an outside task): customer experience **77%**, design **75%**, HR **69%**, legal **56%**, marketing **53%**, sales **40%**, finance **40%**, engineering **28%** — *« a majority in five of eight groups »*. **Two distinct directions of circulation**: design **imports** (35.2%) and barely **exports** anything (1.7%); engineering does the opposite (imports 18.5%, exports 7.4%); **marketing accumulates both** (imports 24.3%, exports **8.9%**, the highest outward share in the sample). **Two tasks appear in the top 3 of borrowings for the other seven groups**: **financial calculation** and **technology troubleshooting**. ⭐ **The heatmap, absent from the coverage, is the richest object**: it gives the full distribution of tasks by user occupation, and its diagonal is striking — engineering retains **53%** of its own work while customer experience retains only **11%**, HR **10%** and design **12%**. **Size effect**: the outside-occupation share drops from **18.9%** (2-5 employees) to **16.3%** (>100 employees) — ⚠️ **but only "among average users"**, OpenAI noting that *« among the heaviest users, we do not see the same monotonic pattern »*, and concluding conditionally: *« AI **may be** especially useful as a generalist tool where specialist resources are scarce. »* **Claimed status**: an **early signal**, visible *« before firms rewrite job descriptions or create new job titles »*. ⚠️ **Structural caveat**: OpenAI measures OpenAI's own usage, on U.S. ChatGPT users only, and presents this position as an asset — *« our unique window into how the world of work is changing »*.

## Titre Article

How AI is expanding what people do at work (Work at the Frontier, rapport 1)

## Date

2026-07-27

## URL

https://openai.com/index/how-ai-is-expanding-what-people-do-at-work/

## Keywords

OpenAI Economic Research, Work at the Frontier, task crossover, task spillover, occupational porosity, 800,000 messages, ChatGPT, U.S. users, work-related messages, occupation-specific messages, generic tasks, calculation base, methodological funnel, inside or near, customer experience, design, human resources, legal, marketing, sales, finance, engineering, task import-export, heatmap, diagonal, occupational retention, financial calculation, technology troubleshooting, marketing material creation, organization size, job positions, average users, heavy users, non-monotonicity, generalist, specialist resources, AI Jobs Transition Framework, occupational reorganization, early signal, job descriptions, job titles, labor-market statistics, division of labor, source bias, vendor-side measurement

## Authors

**OpenAI Economic Research** — équipe de recherche économique d'OpenAI ; la page crédite simplement *« OpenAI »* et la classe sous les tags *Economic Research* et *2026*. Le billet est la porte d'entrée d'un **rapport PDF** (`work-at-the-frontier-report.pdf`) et s'adosse à un cadre antérieur de la même équipe, l'**AI Jobs Transition Framework**, dont il reprend la thèse que de nombreux métiers vont **se réorganiser** plutôt que disparaître.

**Position à connaître** : l'éditeur mesure l'usage de son propre produit et en fait explicitement un argument — *« Using our unique window into how the world of work is changing, we will offer regular data-driven insights based on evidence to guide policy and practice. »* La visée est déclarée : **orienter la politique publique et la pratique**.

## Ton

**Profile**: institutional research post, sober and methodical register, serving as the public summary of a PDF report. Neither a product announcement nor an op-ed — a **findings note** structured in four sections whose headings are the conclusions (*« Nearly half of occupation-specific AI use crosses job boundaries »*, *« Some tasks travel farther than others »*, *« More task crossover in small businesses »*, *« The task list itself is changing »*).

**Style**: **the methodological counter-move stated up front** is the structuring move. *« Many studies of AI and work begin with a fixed list of tasks associated with a given occupation and ask whether models can perform them. Our evidence suggests that AI is also changing who takes on which tasks. »* The object of measurement is not an occupation's substitutability but the **observed redistribution** — a claimed shift in framing, and it is what gives the paper its value.

**Notable trait: caution is built into the sentence, not relegated to a footnote.** Qualifiers are placed at the very moment the result is stated — *« our new research **suggests** »*, *« **Among average users**, the outside-occupation task share falls… »*, *« Among the heaviest users, **we do not see the same monotonic pattern** »*, *« AI **may be** especially useful as a generalist tool »*, *« usage patterns **may** provide an early signal »*. The text almost never asserts without bounding.

**Illustration by concrete case before the figure**: three vignettes open the piece — the small-business owner who drafts copy, reviews a contract, and does a financial analysis; the salesperson who explores a customer dataset *« that might once have gone to an analyst »*; the marketer who fixes a website *« without waiting for a developer »*. The phrase that sums them up is the thesis: *« AI changes not just how work gets done, but who does what. »*

**Marker phrases**: *« task crossover »*, *« who does what »*, *« some activities that once required a handoff can now be done by the person who first encounters the need »*, *« these occupations are "borrowing" these tasks »*, *« before firms rewrite job descriptions or create new job titles »*.

## Pense-betes

- **⭐⭐ The two headline figures, and why both must always be given**: **16.8% of work-related messages** AND **43.5% of occupation-specific messages**. The funnel: **61.5% of work usage is generic** (writing, summarizing, planning — *« shared too broadly across occupations to be evidence of crossover »*) and removed from the calculation; of the remaining **38.5%**, **43.5%** fall outside the occupation. → **43.5% is calculated on 38.5% of usage.** Citing only the 43.5% figure nearly doubles the perceived scale. Both numbers are in the source; usually only one survives in coverage.
- **⭐ The "inside *or near*" nuance**: the complement of the 43.5% is not "within the occupation" but *« inside **or near** a user's occupation »*. The measured boundary is thus **fuzzy by construction**, which is honest and rarely passed on.
- **The methodological counter-move — the paper's real contribution**: the literature starts from a fixed list of tasks per occupation and asks whether the model can perform them; here the question is **who does what**. *« Our evidence suggests that AI is also changing who takes on which tasks. »* → **This measures an observed redistribution, not a theoretical substitutability.** This is what distinguishes this work from exposure studies (cf. the 5-60% ranges discussed in [[dgtresor-ia-effets-emploi-2026-06-30]] and [[sfeir-ia-emploi-risque-decrochage-2026-07-23]]).
- **By occupation, outside-occupation share** (of occupation-specific messages): customer experience **77%** · design **75%** · HR **69%** · legal **56%** · marketing **53%** · sales **40%** · finance **40%** · **engineering 28%**. *« Outside occupation work is a majority in five of eight groups. »*
- **⭐⭐ The heatmap — the richest object, and absent from all coverage** (rows = user's occupation, columns = task's occupation of origin; diagonal in bold): | User ↓ / Task → | CX | Design | Eng. | Fin. | HR | Legal | Mktg | Sales | |---|---|---|---|---|---|---|---|---| | **Customer experience** | **11** | 5 | 20 | 14 | 6 | 6 | **26** | 11 | | **Design** | 6 | **12** | **28** | 10 | 4 | 4 | **28** | 8 | | **Engineering** | 4 | 4 | **53** | 9 | 4 | 4 | 20 | 2 | | **Finance** | 6 | 4 | 22 | **23** | 4 | 8 | **25** | 7 | | **HR** | 9 | 5 | 18 | 16 | **10** | 10 | **23** | 9 | | **Legal** | 5 | 4 | 17 | 12 | 7 | **31** | 18 | 6 | | **Marketing** | 10 | 6 | 17 | 11 | 4 | 4 | **36** | 12 | | **Sales** | 11 | 5 | 18 | 15 | 4 | 6 | **29** | **12** | **Three readings the text does not make**: 1. **The diagonal is the real result.** Engineering retains **53%** of its own work; customer experience **11%**, HR **10%**, design **12%**, sales **12%**. → **Some occupations, in their non-generic use of ChatGPT, have almost stopped doing their own job.** 2. **For four occupations, the marketing column exceeds the diagonal.** A designer does more marketing tasks (28%) and engineering (28%) than design (12%); a customer-experience agent does more marketing (26%) than CX (11%); same for HR (23% vs 10%) and sales (29% vs 12%). 3. **Two columns absorb almost everything**: marketing (18 to 36% everywhere) and engineering (17 to 53%). The other six occupations account for crumbs. ⚠️ **Reading caution**: the low diagonal does not mean these employees no longer do their own job — it means **their non-generic messages** are classified elsewhere, on a base that already excludes 61.5% of usage. The tasks specific to these occupations may be precisely those classified as "generic".
- **The two directions of circulation, quantified**: **design** imports **35.2%** / exports **1.7%**; **engineering** imports **18.5%** / exports **7.4%**; **marketing** imports **24.3%** / exports **8.9%** — *« the highest outward share in the sample »*. → Marketing is the only one to accumulate both directions.
- **The two universally borrowed tasks**: **financial calculation** and **technology troubleshooting** appear in the **top 3 of borrowings for the other seven groups**. **Creating marketing materials** appears in five other groups, *« especially prominent among design users »*.
- **⭐ The asymmetry of engineering — the most actionable figure for an IT services firm**: engineering is **both the least porous occupation (28%, imports 18.5%) and a major exporter (7.4%)**, and its diagonal of **53%** is by far the highest. The celebrated porosity is thus **largely one-way toward the technical side**: other occupations do troubleshooting and technical systems work, the reverse is rare (design exports 1.7%). **What this concretely means**: AI does not dilute engineering into other occupations — it **spreads engineering tasks everywhere else** while leaving engineers on their core work.
- **⚠️⚠️ The size effect is weaker than reported elsewhere — two caveats posed by OpenAI and often lost**: 1. The shift from **18.9% to 16.3%** holds **"among average users"** only. 2. *« **Among the heaviest users, we do not see the same monotonic pattern.** »* → **The trend does not hold for heavy users.** OpenAI offers two possible explanations (moderate users at small organizations turn to AI when they encounter work belonging to another function; heavy users have stabilized workflows that resemble each other across organizations, or use AI more intensively **within** their core occupation). 3. The conclusion is conditional: *« AI **may be** especially useful as a generalist tool where specialist resources are scarce. »* → **Correction to [[sfeir-ia-frontieres-metiers-skill-based-organisation-2026-08-01]]**, which presented these 2.6 points as a clean-cut result and the generalist conclusion as an assertion. The source is more cautious than the commentary on it, and the 2.6-point gap I had flagged as the weak link is **explicitly bounded by the authors themselves**.
- **The claimed epistemic status — early signal**: this data reveals recombinations *« before firms rewrite job descriptions or create new job titles »*, and *« may provide an early signal of occupational change that conventional labor-market statistics will capture only later »*. → **A leading indicator, not an employment measure.** It says what people *attempt*, not what organizations *ratify*.
- **Doctrinal affiliation**: the paper draws on the same team's **AI Jobs Transition Framework**, whose thesis is that many occupations will **reorganize** — *« jobs whose day-to-day tasks could change substantially »* — rather than disappear. Crossover is the usage-level evidence for it.
- **⚠️ Substantive caveats, to be raised systematically**:
- **OpenAI measures OpenAI.** A single assistant, its own users only, presented as an advantage (*« our unique window »*). What is measured is ChatGPT usage, not AI usage.
- **U.S. users only.** No international extrapolation is possible, and the report does not attempt one.
- **Non-representative population** of the workforce: these are ChatGPT users, self-selected.
- **The occupation classification is done by a model**, on messages, with no external validation described in the post. The categories (eight groups) and the "generic / specific" boundary are design choices that mechanically determine the results — this is where the robustness of the 43.5% figure is actually decided.
- **Stated aim**: *« to guide policy and practice »*. An actor that produces the data on its own impact **and** aims to inform public policy. The same bias flagged regarding Bercy in [[sfeir-ia-emploi-risque-decrochage-2026-07-23]].
- **Meta / cross-references**: primary source for [[sfeir-ia-frontieres-metiers-skill-based-organisation-2026-08-01]], which it **supplements** (16.8%, marketing 24.3%, heatmap) and **corrects** (non-monotonic size effect among heavy users); to be read alongside [[dgtresor-ia-effets-emploi-2026-06-30]] and [[sfeir-ia-emploi-risque-decrochage-2026-07-23]] on the gap between theoretical exposure and observed redistribution; extends [[mollick-organizational-theory-agentic-ai-spans-control-2026-02]] and [[mollick-valence-ai-hr-playbook-leader-lab-crowd-2025-07-23]] on the organizational-theory side; HR context in [[bersin-chro-pivotal-role-ai-transformation-2025-10-10]].

## RésuméDe400mots

First installment in **OpenAI Economic Research**'s **Work at the Frontier** series (July 27, 2026), based on more than **800,000 messages** from U.S. ChatGPT users.

**The concept.** ***Task crossover*** refers to *« work historically associated with one occupation appearing in the AI use of people in another »*. The methodological counter-move is set up from the outset: exposure studies start from a fixed list of tasks and ask whether the model can perform them; here the question is **who does what**. *« AI changes not just how work gets done, but who does what. »*

**The figures, and there are two of them.** **16.8%** of work-related messages and **43.5%** of occupation-specific messages concern a task from a different occupation. The gap comes from the funnel: **61.5%** of usage is **generic** (writing, summarizing, planning) and excluded; of the remaining 38.5%, 43.5% fall outside the occupation, the rest being *« inside **or near** »*.

**By occupation**: customer experience **77%**, design **75%**, HR **69%**, legal 56%, marketing 53%, sales and finance 40%, **engineering 28%** — a majority in five of eight groups.

**Two directions of circulation.** Design **imports** (35.2%) without exporting (1.7%); engineering does the opposite (18.5% / 7.4%); marketing **accumulates both** (24.3% / 8.9%, the highest outward share). Two tasks appear in the top 3 of borrowings for the other seven groups: **financial calculation** and **technology troubleshooting**.

**The heatmap** gives the full distribution, and its diagonal is the most striking result: engineering retains **53%** of its own work, while customer experience retains only **11%**, HR **10%** and design **12%** — for these three occupations, marketing tasks outweigh their own.

**The size effect is more fragile than it appears.** The outside-occupation share drops from 18.9% (2-5 employees) to 16.3% (>100 employees) **among average users only**: *« among the heaviest users, we do not see the same monotonic pattern »*. The conclusion remains conditional — *« AI **may be** especially useful as a generalist tool where specialist resources are scarce »*.

**The claimed status** is that of an **early signal**, visible *« before firms rewrite job descriptions or create new job titles »*.

⚠️ OpenAI measures the usage of its own product, on its U.S. users only, and presents this position as an asset.

## GrapheDeConnaissance

- OpenAI Economic Research —publie→ Work at the Frontier (DOCUMENT, 0.98)
- Work at the Frontier —mesure→ 16,8 % des messages liés au travail et 43,5 % des messages métier-spécifiques portent sur une tâche d'un autre métier (MESURE, 0.97)
- Work at the Frontier —mesure→ 61,5 % de l'usage professionnel est générique et écarté du calcul du crossover (MESURE, 0.95)
- task crossover —est_instance_de→ l'apparition, dans l'usage IA d'un métier, de travail historiquement associé à un autre (CITATION, 0.97)
- OpenAI Economic Research —s_oppose_à→ les études qui partent d'une liste figée de tâches par métier pour tester la substituabilité du modèle (AFFIRMATION, 0.93)
- Work at the Frontier —mesure→ part hors métier par fonction : expérience client 77 %, design 75 %, RH 69 %, juridique 56 %, marketing 53 %, vente 40 %, finance 40 %, ingénierie 28 % (MESURE, 0.96)
- Work at the Frontier —mesure→ le design importe 35,2 % de tâches extérieures et n'en exporte que 1,7 % (MESURE, 0.95)
- Work at the Frontier —mesure→ l'ingénierie importe 18,5 % de tâches extérieures et en exporte 7,4 % (MESURE, 0.95)
- Work at the Frontier —mesure→ le marketing importe 24,3 % et exporte 8,9 %, la plus forte part sortante de l'échantillon (MESURE, 0.95)
- Work at the Frontier —mesure→ l'ingénierie retient 53 % de ses propres tâches, contre 11 % pour l'expérience client et 10 % pour les ressources humaines (MESURE, 0.92)
- calcul financier —observé_dans→ le top 3 des tâches empruntées par les sept autres groupes de métiers (AFFIRMATION, 0.94)
- dépannage technologique —observé_dans→ le top 3 des tâches empruntées par les sept autres groupes de métiers (AFFIRMATION, 0.94)
- Work at the Frontier —mesure→ la part hors métier passe de 18,9 % à 16,3 % selon la taille de l'espace de travail, chez les utilisateurs moyens seulement (MESURE, 0.95)
- OpenAI Economic Research —affirme_que→ la tendance liée à la taille de l'organisation n'est pas monotone chez les utilisateurs les plus intensifs (CITATION, 0.95)
- OpenAI Economic Research —affirme_que→ l'IA pourrait être particulièrement utile comme outil généraliste là où les ressources spécialisées manquent (AFFIRMATION, 0.9)
- données d'usage —permet→ de voir la recomposition des métiers avant la réécriture des fiches de poste et des intitulés (AFFIRMATION, 0.94)
- Work at the Frontier —est_basé_sur→ AI Jobs Transition Framework (DOCUMENT, 0.92)
- AI Jobs Transition Framework —affirme_que→ de nombreux métiers vont se réorganiser, leurs tâches quotidiennes pouvant changer substantiellement (AFFIRMATION, 0.92)
- Work at the Frontier —est_basé_sur→ plus de 800 000 messages d'utilisateurs américains de ChatGPT (AFFIRMATION, 0.97)
- OpenAI —affirme_que→ sa position d'observation unique sur le monde du travail lui permet d'orienter la politique publique et la pratique (CITATION, 0.9)

---
Canonical: https://www.thekb.eu/en/fiches/openai-work-at-the-frontier-task-crossover-2026-07-27/