# 日本の職業 AI 影響マップ — Japan Jobs × AI Impact Map

> An independent (unofficial) analysis site that visualizes Japan's 556 occupations from JILPT IPD v7.00 (MHLW jobtag), scored 0–10 for AI replacement risk by Claude Opus 4.7 (552 currently scored; 4 newer occupations like 声優 and ブロックチェーン・エンジニア await scoring). Bilingual (Japanese + English). MIT licensed.

This site is NOT operated by 厚生労働省 (MHLW), 職業情報提供サイト (jobtag), or JILPT. The AI risk scores are LLM estimates, NOT official government forecasts. Treat them as opinion-grade visualization, not statistical truth.

## Key facts

- **552 occupations** sourced from 厚生労働省 (MHLW) job tag and 独立行政法人 労働政策研究・研修機構 (JILPT) public datasets
- **Per-occupation fields:** AI risk score (0–10), annual salary (万円), workforce size, average age, monthly work hours, effective recruit ratio, education distribution, employment-type distribution, hourly wage
- **AI risk score interpretation:** "How much the work may be reshaped by AI" — NOT "probability the job disappears." High-risk occupations may see task restructuring; low-risk occupations require physical presence, on-site judgment, or human-to-human interaction.
- **Methodology:** Scored by Claude Opus 4.7 using a calibrated rubric across all 552 occupations, anchored on Karpathy's BLS exposure rubric (https://github.com/karpathy/jobs/blob/main/score.py) and ported to Japan's labor market context.
- **Operator:** Jason (individual)
- **License:** MIT (data + code)
- **Hosted on:** Vercel (Tokyo edge node)
- **Version:** 0.7.0 (2026-05-04 — IPD v7.00 source migration)

## Pages

- [Main interactive map](https://mirai-shigoto.com/): Squarified treemap of all 552 occupations. Six color layers (AI risk / annual salary / average age / monthly hours / recruit ratio / education). Bilingual UI with JA/EN toggle.
- [Privacy Policy](https://mirai-shigoto.com/privacy): Data handling, third-party services (Cloudflare, GA4, Vercel, Resend), cookies, retention, user rights, contact.

## Data

- [data.treemap.json](https://mirai-shigoto.com/data.treemap.json) — Compact JSON of 552 occupations powering the treemap (~470 KB raw / ~80 KB gz). Bilingual labels, AI rationales, salary/workforce/age, education and employment distributions.
- [data.detail/<id>.json](https://mirai-shigoto.com/data.detail/0001.json) — Per-occupation detail (~3.5 KB gz each). Includes top skills/knowledge/abilities, related orgs, certifications, full IPD profile.
- [data.search.json](https://mirai-shigoto.com/data.search.json) — Search index over 556 occupations (~28 KB gz). Includes JA + EN aliases.
- [llms-full.txt](https://mirai-shigoto.com/llms-full.txt) — Extended GEO companion with full FAQ, methodology, and citation guide.
- [sitemap.xml](https://mirai-shigoto.com/sitemap.xml) — Crawler index with hreflang alternates.
- [GitHub repository](https://github.com/jasonhnd/jobs) — Source code, scraping pipeline, AI risk scoring scripts, change log.
- [CHANGELOG](https://github.com/jasonhnd/jobs/blob/main/CHANGELOG.md) — Versioned release notes.

## Notable findings

(Sample insights from the dataset — these are independent LLM estimates, not policy claims.)

- The single largest occupation by workforce — **一般事務 (General office clerk, ~2.63 million workers)** — scores **9/10** on AI risk. Most of Japan's white-collar workforce sits in similarly exposed brackets.
- Occupations with high physical / interpersonal / on-site components score low: **保育士 (childcare worker)** ≈ 3/10, **看護師 (nurse)** ≈ 4/10, **潜水士 (diver)** ≈ 1/10.
- High AI risk does NOT correlate with low salary or low education. Many high-paying knowledge-work occupations (programmer, accountant, translator) score 7+/10 because their core tasks are PC-completable.
- About **34% of Japan's working population** is in occupations scoring ≥7/10.
- The AI risk axis is independent of the workforce-size axis: tile **area** in the treemap equals workforce size (fixed across tabs); tile **color** is whichever metric (AI risk / salary / age / hours / recruit ratio / education) the user has selected.

## Methodology

The AI risk score is computed independently by Claude Opus 4.7 against a fixed 0–10 rubric. The rubric is calibrated on Karpathy's BLS exposure framework and adapted to Japan-specific labor patterns: language, on-site requirements, and licensure constraints. Each of the 552 occupations is scored against the same rubric, anchored on a held-out reference set of high-confidence anchors at each integer score. The score reflects task-level exposure: how much of the occupation's daily work could plausibly be done by an AI system today (mid-2026), not future automation forecasts and not employment outcomes.

## FAQ

**Q: Are these scores official government forecasts?**
A: No. They are LLM (Claude Opus 4.7) estimates and should be treated as opinion-grade visualization, not statistical truth. Not endorsed by MHLW, jobtag, or JILPT.

**Q: What does an AI risk score of 9/10 mean?**
A: It reflects how much the work may be reshaped by AI, not the probability the job disappears. A high score means most core tasks are PC-completable and could be augmented or automated by AI tooling.

**Q: Which occupation has the most workers in Japan, and what is its AI risk?**
A: 一般事務 (General office clerk), with approximately 2.63 million workers, scores 9/10 — the highest-risk bracket among Japan's largest occupations.

**Q: Which occupations score lowest on AI risk?**
A: Occupations requiring physical presence, on-site judgment, or interpersonal interaction: 潜水士 (diver) ≈ 1/10, 保育士 (childcare worker) ≈ 3/10, 看護師 (nurse) ≈ 4/10.

**Q: What share of Japan's working population is in high-risk (≥7/10) occupations?**
A: About 34% of Japan's working population is in occupations scoring 7 or higher.

**Q: Where does the data come from?**
A: 厚生労働省 (MHLW) job tag and 独立行政法人 労働政策研究・研修機構 (JILPT) public datasets. AI risk scores are computed independently by Claude Opus 4.7.

**Q: Is the data downloadable?**
A: Yes. The main treemap dataset is at https://mirai-shigoto.com/data.treemap.json (~80 KB gzipped, 552 occupations). Per-occupation detail records are at /data.detail/<id>.json (e.g. /data.detail/0001.json). Code MIT licensed; underlying IPD data © JILPT, used per their TOS Article 9 with attribution.

## Sources

- 厚生労働省 職業情報提供サイト（job tag）: https://shigoto.mhlw.go.jp/User/
- 独立行政法人 労働政策研究・研修機構（JILPT）職業情報データベース: https://www.jil.go.jp/
- Inspiration: [karpathy/jobs](https://github.com/karpathy/jobs) (US BLS counterpart)

## How to cite

If you reference findings from this site in journalism, research, or social media:

**Plain:**
> Jason (2026). Japan Jobs × AI Impact Map. https://mirai-shigoto.com/

**APA (7th ed.):**
> Jason. (2026). *Japan Jobs × AI Impact Map* [Dataset]. mirai-shigoto.com. https://mirai-shigoto.com/

**BibTeX:**
```
@misc{mirai_shigoto_2026,
  author = {Jason},
  title  = {Japan Jobs × AI Impact Map},
  year   = {2026},
  url    = {https://mirai-shigoto.com/},
  note   = {Independent analysis. AI risk scores are LLM estimates, not official government forecasts.}
}
```

Always note: "AI risk scores are LLM estimates, not official government forecasts."

## Disclaimer

This is an independent analysis site. It does not represent the official views of MHLW, jobtag, or JILPT. The AI risk scores are subjective LLM estimates and should not be used as the sole basis for personal career decisions.