🚩 Report: Copyright infringement
I am writing to formally notify you that the dataset you have published/uploaded contains my copyrighted creative works, as well as those of other European creators who post on this site, without our explicit consent.
This dataset is being distributed and potentially utilized for training Artificial Intelligence models in direct violation of European Union law.
Please be advised of the following legal obligations and non-compliance issues:
Violation of the EU AI Act (Regulation (EU) 2024/1689): Under Article 53 of the EU AI Act, developers of AI models and providers of AI datasets are strictly mandated to implement policies that guarantee compliance with European Union copyright laws and technological restrictions. The platform from which you harvested this data explicitly deploys machine-readable protections to prohibit scraping and AI training. By systematically bypassing these technical protocols, organizing data into a structured .db file, and making it publicly available, you are actively facilitating non-compliance with the EU AI Act.
European artists' unauthorized inclusion in this AI-targeted training set is definitively proven by your database records containing unique Author IDs and specific Post IDs.
Infringement of the EU Copyright Directive (Directive (EU) 2019/790 - DSM): Our artistic works are protected under EU copyright frameworks. In strict accordance with Article 4(3) of the DSM Directive, We have explicitly exercised our right to opt-out of text and data mining (TDM) through platform-wide configurations and machine-readable 'NoAI' metadata. Your retention and publication of operational Post IDs and platform shortcuts serve as an illegal index for copyright extraction. Compiling, hosting, and distributing this tailored structural blueprint to bypass our TDM reservation constitutes a direct infringement of our copyright protections under Union law.
Compliance Breach under the General Data Protection Regulation (GDPR - Regulation (EU) 2016/679): The published dataset illegally processes and exposes our permanent platform-specific identifiers, specifically our unique Author IDs and associated Post IDs. Also the dataset indexes and replicates our visible artistic signature and pseudonym embedded within our creative works. Under EU law, these elements combined constitute personal data, as they form a distinct, irreversible digital footprint that links directly back to my real-world identity and professional profile. Processing, profiling, and publicly redistributing this structured identity index for AI training purposes without a valid lawful basis under Article 6 of the GDPR is an explicit violation.
So, I demand that you (and other people who download this database for training purposes) immediately:
- Remove database , metadata, and associated personal data from your dataset.
- Stop and desist from any further distribution, hosting, or sharing of this dataset on any public or private repository (including but not limited to Hugging Face, GitHub, or cloud drives).
- Confirm in writing once the removal has been completed, publish a formal apology to the artists and the owners of the Cara platform, and guarantee that you will respect and actively promote the opt-out principle and website security against unauthorized scraping within your community in the future.
Failure to comply with this request immediately will leave us with no choice but to escalate this matter.
I count on your cooperation.
You have 48 hours to delete the databases and copies, otherwise I will start working on bringing this matter to the appropriate institutions.
The platform from which you harvested this data explicitly deploys machine-readable protections to prohibit scraping and AI training. By systematically bypassing these technical protocols, organizing data into a structured .db file, and making it publicly available, you are actively facilitating non-compliance with the EU AI Act.
This is false—there was zero bypassing of technical protocols. I encountered exactly zero roadblocks while building this dataset. In fact, my scripts didn't even have to emulate a browser session. Such scripts would not have worked on any serious website like X, the everything app (formerly Twitter), or Instagram.
Infringement of the EU Copyright Directive (Directive (EU) 2019/790 - DSM): Our artistic works are protected under EU copyright frameworks
This dataset contains no image data. It is an index dataset; search engines do the exact same thing.
publish a formal apology to the artists and the owners of the Cara platform, and guarantee that you will respect and actively promote the opt-out principle and website security against unauthorized scraping within your community in the future
Hefty—who actually did download 12 TB worth of data—did exactly that, and then, if my memory doesn't betray me, refused to sign an NDA (to help cara.app directly) because he was too spooked by the number of death and doxxing threats he received from your community.
will leave us with no choice but to escalate this matter.
who is "us" in this context?
First please stop bluffing, you have no idea how complicated an international lawsuit against some dude online is. Secondly, you never stopped sending death threats to the first guy so, why should he care?
Also this data set will help the open source ai community, which everyone can use locally on their pc, it's not like he is sending it to open ai or anthropic.