FranceCPI Article L122-5-3, Text and Data Mining Exception
archived copy
Read from a public archive copy, not the publisher directly. The publisher does not serve this page to automated readers, so a direct fetch was not an option here; how we access sources. Archived 2024. Publisher's page: https://www.legifrance.gouv.fr/codes/article_lc/LEGIARTI000044363192In force since .
A text and data mining (TDM) opt-out rule binding public and private bodies.
- Instrument type
- an act of a legislature
What it requires
- Text and data mining means applying an automated analysis technique to digital text and data to extract information such as patterns, trends and correlations.
- You may reproduce a lawfully accessed work for text and data mining, whatever its purpose, without the author's authorization, unless the author has appropriately objected, including by machine-readable means for content made available online (para. III).
- Before mining, check for that objection; an indexing or training use that ignores a machine-readable opt-out falls outside the exception and requires authorization.
- If you are a research organization, public-access library, museum, archive or film, audiovisual or sound heritage institution, or act on their behalf and at their request, you may reproduce a lawfully accessed work for text and data mining carried out solely for scientific research, without authorization (para. II).
These lines are LexLint's reading of the law's text.
In the frameworks
- NIST AI RMF
- MAP 4.1Approaches for mapping AI technology and legal risks of its components – including the use of third-party data or software – are in place, followed, and documented, as are risks of infringement of a third party’s intellectual property or other rights.
Our reading of where this law's duties sit in each framework. It never says that running a control meets the law. How we read a framework against the law
Who enforces it
Settledness
- As of
- Open questions
- Does a reservation that a site states only in its terms of use count as an objection made 'de manière appropriée' under Article L122-5-3 CPI, or must content made available online carry it by 'procédés lisibles par machine'?
What this law does
Article L122-5-3.I defines text and data mining as implementing an automated analysis technique on digital text and data to extract information such as patterns, trends, and correlations, for the purposes of Article L122-5's 10° exception.
Paragraph II lets research organizations, public-access libraries, museums, archives, and film, audiovisual, or sound heritage institutions, or others acting on their behalf and at their request including under a nonprofit partnership with private actors, reproduce a lawfully accessed work for text and data mining carried out solely for scientific research, without the author's authorization.
Paragraph III separately lets any person reproduce a lawfully accessed work for text and data mining regardless of its purpose, unless the author has appropriately objected, including by machine-readable means, for content made available online. A use that mines a work whose rightholder has expressed a machine-readable opt-out under paragraph III falls outside the exception and requires authorization.
When LexLint raises it
When your app profile says your app reuses other publishers' content, crawls the web or trains models.
Back to the example · Lint your app
Page updated .