Every AI conversation on Earth falls into one of three buckets: work, personal life, or study.
Measure that split across 121 countries — as Anthropic's Economic Index does — and you get something unexpected: a map of national character, drawn in prompts.
Americans, it turns out, mostly ask AI about their personal lives. Brazilians put it to work. Tunisians hand it their homework.
And buried in this apparently charming dataset is one of the most useful signals an international creditor can have: which countries' businesses are actually tooled — regardless of what headline adoption says.
The work/life board
Global baseline: 43% work, 40% personal, 16% coursework. Now look at who deviates.
STATION REPORT · USE-CASE SPLIT · MAY 2026
Work / Personal / Study — Selected Countries
Share of each country's AI conversations by use case. Source: Anthropic Economic Index, May 2026 — a snapshot of conversation content, not a survey of users.
Three portraits from the board
Brazil: the world's most work-obsessed AI profile. 57% of Brazilian usage is work — the highest share of any major economy, fourteen points above the global norm. Remember that Brazil ranks a modest 61st in overall adoption. Put those together and the picture sharpens: fewer Brazilians use AI, but the ones who do are overwhelmingly using it in a business. The Gulf tells the same story — the UAE (55%) and Qatar (51%) both run heavily work-weighted profiles on mid-table adoption.
America: the machine is for life, not the office. The US and Canada are near mirror images of Brazil — roughly 50% personal usage against 41–42% work. The richest AI markets on Earth spend their prompts on hobbies, health, plans and feelings. Germany, France, Iceland and Finland cluster the same way. Whatever this says about civilisation, it says something concrete about offices: high national adoption does not mean the finance department is tooled.
Tunisia: the classroom republic. 52% of Tunisian usage is coursework — the highest study share on Earth, with Algeria (47%), Indonesia (46%) and Bolivia (43%) close behind. An entire generation across North Africa, Southeast Asia and the Andes is doing its homework with a machine. Check back on those countries' back offices in ten years.
The creditor's takeaway: read the work share, not the headline
Here's why this board earns its place in a collection agency's journal.
In our previous Station Report, we tiered countries by headline adoption. The work/life split is the correction layer on top of it:
CORRECTION LAYER · WORK SHARE VS HEADLINE ADOPTION
How to Read a Debtor Country in Two Numbers
The single-sentence version: the work share tells you whether the machine sits in the office or on the sofa. For predicting what answers your payment reminder, the office is the only number that matters.
Honest footnotes
As always with this dataset: it measures conversation content in a single-period snapshot — not users, not employment, and with no trend line. Deviations describe behaviour, not destiny. And per-country profiles describe averages; your specific debtor may be the exception in either direction. That last point isn't a weakness of the data — it's the reason collection remains a case-by-case craft.
Frequently asked questions
Which country uses AI for work the most?
Brazil leads major economies with 57% of its AI usage being work-related, followed by the UAE (55%), Israel and Vietnam (53%), and Thailand (53%). The global average is 43%.
Which countries use AI mostly for personal life?
The United States and Canada lead, with roughly 50% of usage being personal — above their work shares. Germany, France, Iceland and Finland show the same personal-leaning pattern.
Which country uses AI most for studying?
Tunisia, where coursework accounts for about 52% of all usage — the highest share measured — followed by Algeria (47%), Indonesia (46%) and Bolivia (43%).
Why does the work/personal AI split matter for international trade and collection?
Because it corrects the headline adoption numbers. A country with modest overall adoption but a high work share (like Brazil or the UAE) hosts businesses that are likely tooled — meaning machine-drafted commercial correspondence — while a high-adoption country with a personal-leaning profile offers no such guarantee about any given company.
Every jurisdiction has a personality. We've met them all.
The data now confirms what thirty years of cross-border collection taught us the slow way: countries don't just have different laws. They have different relationships with work itself.
InterStation collects B2B debts worldwide — with local teams who knew their market's character long before it was measurable.
Related reading: AI Adoption by Country 2026 · Why Money Hubs Adopted AI First