Ministers rule out giving Estonian content to AI companies for free

Estonian Prime Minister Kristen Michal and Minister of Culture Heidy Purga both ruled out the possibility of handing over Estonian-language data, including media content belonging to ERR, to large artificial intelligence (AI) companies for free. The idea had been backed by Minister of Justice and Digital Affairs Lisa Pakosta (Eesti 200).
"The material that has been created in the media and in various repositories – it certainly should not be given away lightly and for free," Michal said at a government press conference on Thursday.
"Yesterday, we met with the editors-in-chief of regional newspapers and major publications, and this was one of the main topics on the agenda – that the material that has been collected, in archives and in the editorial offices of various newspapers and publications, is a repository that has been created over time and does not have to be given away for free to the first person who asks for it," he said.
According to Michal, the Estonian government is on the same side as the media in this matter and values the material they have created, which is of high quality and great importance. "It certainly must not be given to the first person who asks for it and for practically just a handshake," he stressed.

Minister of Culture Heidy Purga (Reform) also said that in her view, Estonian databases, including the data held by ERR, are worth their weight in gold and should not be handed over lightly to anyone, though it depends on negotiations, and the conditions.
"Although the direction is undoubtedly the right one – it is certainly important for Estonia culturally, economically and definitely from a security perspective too, to have the capacity to create and train autonomous models. And where else can we get this standard from, when including the Estonian Language Institute and ERR is very important in terms of the quality criteria for this standard. Here we have to look toward copyright, licensing and other agreements," Purga said.
"And if I may interject with a rhetorical aside here, Meta (Facebook's parent company – ed.) does business in our country without paying taxes. We also have to take this into account," Purga added.
Andres Jõesaar, media adviser at the Ministry of Culture, commented on Justice and Digital Affairs Minister Liisa Pakosta's approach on social media. He said the survival of the Estonian language and culture does not depend on Meta or X, but on Estonia's own ability to handle AI and develop large language models.
"It's one thing to just share words, or a corpus of language, that can be considered, but it's quite another to share content – articles, books, news, films and anything else created with those words. The idea that we have to give away all the content created by authors to the big players for free sounds rather like DT's (Donald Trump – ed.) insistence that Ukraine must give away its precious mineral resources to the U.S. in exchange for its freedom. Neither deal is a good deal."

Minister of Education and Research Kristina Kallas (Eesti 200) said this is a really important issue for the future of the Estonian language, as the corpus currently available in Estonian for use in large language models is very limited.
"This means that if the Ministry of Education plans to launch a project using artificial intelligence in Estonian educational institutions from the fall, the big question whether the artificial intelligence we employ can actually use Estonian," she said.
According to Kallas, AI language models are already using Estonian, but on the basis of the very limited language material they have available.
"A very large proportion of the corpus of Estonian is not available to large AI systems today. ERR is one thing, but quite another is the language corpus of the Estonian Language Institute, which is not available, nor is the Estonian-language research that is covered by the license – master's theses, bachelor's theses, doctoral theses," Kallas said.
"If we don't want Estonian to become a language that disappears from technology as technology develops, that is that it doesn't use Estonian, and is not capable of developing further – and we don't want that – then I think it is absolutely the right decision to make it available to major language models as well," she continued.
The education minister also stressed however that this should not be done without restrictions, but in a legally correct way.
"Making the Estonian language available to large language models does not mean we make it available to everyone without any restrictions. However, it is a question of agreements – to what extent this can be done. There are legal issues regarding copyright, there are legal issues with personal data protection and these must first be resolved legally," Kallas said.
"But if you ask me in principle whether Estonia should move in the direction of making the Estonian language corpus available, the answer is yes, because otherwise there is a risk that the Estonian language will become much weaker globally," she said.
On Thursday, Kallas' party colleague, Minister of Justice and Digital Affairs Lisa Pakosta (Eesti 200), told ERR that giving Estonian-language data, including ERR's Estonian-language media content, free of charge to large artificial intelligence companies would contribute to ensuring the survival of the Estonian language.
---
Follow ERR News on Facebook and Twitter and never miss an update!
Editor: Mait Ots, Michael Cole