Minister of Justice and Digital Affairs Liisa Pakosta said that with the amendment, all public texts without an opt-out mark could be readable by artificial intelligence. The "thinking" of AI is considered here to be similar to that of scientists.
Pakosta said that the issue is how to clarify the legal framework for the use of currently available data. She noted that 13 years ago, when the Estonian Language Institute began developing a language model, the legal foundations should have been addressed more clearly.
On Monday, the minister, representatives of ministries, and scientists discussed in the Riigikogu Cultural Committee the role that language corpora play in the development of artificial intelligence. They talked about what to do so that AI would speak better Estonian in the future.
Chair of the Cultural Committee Liina Kersna said that Estonia has a constitutional obligation to ensure that artificial intelligence speaks the best possible Estonian. She emphasized that this is very important for the survival of the language.
According to Kersna, there are two options on the table: first, the development of a language corpus that AI developers could use, and second, a domestic or European collaborative AI platform. This would allow the state to use its data more for the development of new services.
Kersna emphasized that the higher the quality of the text for training AI, the better it will communicate. Therefore, it is important that the input material is of good quality.
She added that lawyers have not found a point in the law that would prohibit AI from using public texts. The Ministry of Justice has sent a draft proposal for consultation, stating that public language data may be used by AI for scientific and development activities.
It is not yet clear when the legislative amendment will reach the government or the Riigikogu.