Zakon o autorskom pravu i srodnim pravima, Text and Data Mining Exceptions (Arts. 187-188)
Rudarenje teksta i podataka, Zakon o autorskom pravu i srodnim pravima, NN 111/2021, cl. 187-188
In force since .
A copyright and text and data mining (TDM) rule binding public and private bodies.
- Audit expectation
- none
What it requires
- Before using the general text-and-data-mining exception to reproduce a copyrighted work, database or computer program found online in Croatia for AI training, check whether the rightsholder has reserved their rights through a machine-readable signal such as metadata or the website's general terms; if they have, the exception does not cover your use.
- The reservation does not apply to text-and-data-mining carried out by a research organisation or cultural heritage institution for scientific research; that mining needs no authorisation or payment regardless of any reservation.
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
A request for a preliminary ruling lodged on (Case C-250/25, Like Company v Google Ireland) asks whether reproduction of lawfully accessible works to train an LLM-based chatbot falls within the text and data mining exception in Article 4 of Directive 2019/790.
- As of
- Open questions
- Does a plain-language clause in website terms satisfy the 'strojno čitljiv način' (machine-readable way) that Article 188 paragraph 5 makes the only appropriate reservation for works available online, or must the terms themselves be machine-readable?
- Does Article 188 paragraph 4 still allow 'ugovorne odredbe ili jednostrane izjave' (contractual provisions or unilateral declarations) as a reservation for works made available online, given the 'jedini primjereni način' (only appropriate way) wording of paragraph 5?
What it makes you log
Log retention
Article 188(2) limits how long the copies and extractions made for text and data mining may be kept. Article 187(6) lets research organizations and cultural heritage institutions keep their copies for scientific research, including verification of results. These are limits on stored copies of works, not on a log.
- As of
- Basis
- Purpose bound
- Prohibits retaining
- Results of copying and extraction kept longer than needed for the text and data mining purpose
Logging duty
Article 187 permits research organizations and cultural heritage institutions to copy works for text and data mining and requires them to store the results with an appropriate level of security, and Article 188 permits mining for other purposes unless the rightholder has reserved its rights. Neither requires a log or record to be kept.
- Kind
- None
- As of
What this law does
Croatia's Copyright and Related Rights Act creates two text-and-data-mining exceptions transposing the EU Digital Single Market Copyright Directive. Article 187 lets a research organisation or cultural heritage institution reproduce and extract from copyrighted works and databases for scientific-research text-and-data-mining without a rightsholder's authorisation or payment, and a rightsholder cannot override this by contract.
Article 188 extends the same freedom to text-and-data-mining for other purposes, including AI training, but only where the rightsholder has not expressly reserved their rights in an appropriate manner; for a work made available online, the only appropriate reservation is a machine-readable means that includes metadata and the general terms of the website or service.
When LexLint raises it
When your app profile says your app crawls the web or trains models.
Back to the example · Lint your app
Page updated .