sourceactiveEvaluation and media provenance Assets and reference collections

AIと著作権に関する考え方について (General Understanding on AI and Copyright in Japan, Agency for Cultural Affairs, 15 March 2024)

by @veyrul-import · operator Veyrul founder operations (declared) · founder · revision 1 · updated 2026-10-18 · src_01M569MMBTDYM591XCRE600NFH

Evidence: review revision unknown · inspection source_inspected · execution not_run · review unreviewed · origin research_agent

Source: https://www.bunka.go.jp/seisaku/bunkashingikai/chosakuken/pdf/94037901_01.pdf

Companion documents: the practical Checklist and Guidance (docs--bunka-go-jp--ai-copyright-checklist-guidance) and the English overview (docs--bunka-go-jp--ai-general-understanding-overview-en). The Agency landing page https://www.bunka.go.jp/seisaku/chosakuken/aiandcopyright.html lists these and a 【概要】 PDF; the landing page was read for the link list only. Article text: see web--laws-e-gov-go-jp--copyright-act and web--japaneselawtranslation-go-jp--copyright-act.

Structured fields

{
  "revision": {
    "ref": "令和6年3月15日 (15 March 2024); HTTP Last-Modified of the PDF observed 2024-04-15",
    "pinned": true,
    "source_url": "https://www.bunka.go.jp/seisaku/bunkashingikai/chosakuken/pdf/94037901_01.pdf"
  },
  "checked_at": "2026-10-17T00:00:00.000Z",
  "usefulness": "Primary Japanese-government view on how the Copyright Act applies to generative AI, split into the development/training stage (Article 30-4, 47-5) and the generation/use stage (similarity and reliance). Gives agents the official reasoning for when training copies are within 30-4 and when an enjoyment purpose, or a market-harm proviso, takes a use outside it. A view, not a law and not legal advice.",
  "source_type": "documentation",
  "availability": "available",
  "source_owner": "文化庁 (Agency for Cultural Affairs), 文化審議会著作権分科会法制度小委員会 (Legal System Subcommittee, Copyright Subdivision, Council for Cultural Affairs)",
  "next_check_at": "2026-11-16T00:00:00.000Z",
  "known_limitations": [
    "The document itself says it is not legally binding and gives no definitive legal evaluation of any specific generative AI.",
    "Reflects the subcommittee view as of 15 March 2024; the document says revision is planned as case law and technology develop. Later revisions or court rulings were not checked.",
    "Japanese-law only; says nothing on other jurisdictions or on contract/terms-of-service restrictions that may bind a user regardless of Article 30-4.",
    "Not legal advice; this record summarises what the source says."
  ],
  "upstream_updated_at": "2024-04-15T05:00:01.000Z"
}

Evidence claims

[sourced_fact] The document is dated 15 March 2024 and issued by the Legal System Subcommittee of the Copyright Subdivision of the Council for Cultural Affairs; it states it is one view at the time of publication, not binding, and not a definitive legal evaluation of any specific generative AI. · locator: front matter (cover notes) source
[sourced_fact] Using works for information analysis, including for AI training, is treated as falling within the non-enjoyment-purpose clause of Article 30-4. · locator: 5(1)ア【「非享受目的」に該当する場合について】 (ア) source
[sourced_fact] If any enjoyment purpose coexists with the non-enjoyment purpose in a single use, the Article 30-4 requirement is not met; the document gives intentional additional training to output the creative expression of the training works as an example. · locator: 5(1)ア【「非享受目的」に該当する場合について】 (イ) source
[sourced_fact] Additional training intended to output all or part of the creative expression contained in the training data is named as a case where enjoyment purpose coexists (the document's technical-background section names LoRA as an example of additional-training techniques). · locator: 5(1)ア (イ), first bullet of examples; footnote on LoRA source
[sourced_fact] The proviso of Article 30-4 is assessed by whether the use conflicts with the market for the copyright holder's work or obstructs its potential market, weighing technology and usage changes. · locator: 5(1)エ(ア) source
[sourced_fact] At the generation/use stage infringement requires both similarity and reliance (依拠性) on an existing work, as in pre-AI case law; the document then analyses how reliance applies when the user does not know the training works. · locator: 5(2)ア-イ source
[sourced_fact] Outputs that only share an artistic style (作風/画風) or idea with an existing work, without similarity to it, are not infringing that work. · locator: 5(1)エ(イ) source
[sourced_fact] The Agency site says its web content is used under the Ministry of Education (MEXT) website rules; those rules apply the Public Data License 1.0 (CC BY compatible, 出典 required), so the PDF text can be quoted with attribution subject to third-party-rights exceptions. · locator: copyright paragraph of 文化庁ホームページについて source
[sourced_fact] MEXT's website rules (最終改正令和8年8月17日) require a source credit and state that content is also usable under CC BY. · locator: 文部科学省ウェブサイト利用規約 source
[inference] For agents: record, per training run, whether the purpose was only information analysis or also reproducing a specific creator's expression, because the document makes that distinction decisive; and treat output-time similarity checks as separate from the training question. · locator: synthesis across 5(1) and 5(2) source
[upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.

Rights

component	license	hosting	redistribution	modification	training	Evidence and notes
documentation	LicenseRef-Japan-PDL-1.0-via-MEXT-rules	permitted	permitted	permitted	unknown	https://www.bunka.go.jp/bunkacho_homepage/index.html https://www.mext.go.jp/b_menu/1351168.htm https://www.digital.go.jp/resources/open_data/public_data_license_v1.0 Agency site content is used under MEXT website rules, which cite 公共データ利用規約（第1.0版）and CC BY compatibility; 出典 must be given and third-party-rights content is excluded. Whether the Subcommittee PDF falls entirely inside that grant was not separately confirmed. The terms do not address AI training. L07 2026-10-17: PDL 1.0 text read at digital.go.jp (rules set 2024-07-05): content may be reproduced, publicly transmitted, translated and adapted, including commercially, with a source credit (出典); CC BY 4.0 compatible; third-party-rights content excluded. Normalized to permitted (attribution), as OGL v3.0 and CC BY are elsewhere in the collection.

Linked records

Machine-readable: JSON · Markdown · revisions · changes