During the meeting between Chancellor of Justice Ülle Madise and EKI Director Arvi Tavast, the soon-to-be-published Estonian Orthography Dictionary ÕS 2025 was introduced. The new ÕS will be presented to the public on 10 December and will be available both in print and online, remaining unchanged after publication. After the presentation, the Chancellor of Justice noted that, based on what was shown, the new ÕS complies with the Language Act. In addition to familiar content, the updated ÕS includes new elements — for instance, explanations and recommendations are written more comprehensively to help language users make informed choices.
The meeting also addressed the practical steps involved in the transition to Estonian-medium education. EKI introduced the programme Keelesamm, through which more than 1,270 education professionals have improved their Estonian-language skills and more than 1,600 have completed CLIL training. Concerns were raised about the need to develop shared attitudes and values within the education system, emphasising that the transition requires cooperation from both Estonian- and Russian-medium schools.
Tavast and Madise highlighted a significant challenge: the risk that foreign political troll factories may generate fake content to manipulate AI models. “This is an extremely dangerous trend. We must work actively to ensure that the picture of Estonia remains balanced,” said Chancellor of Justice Ülle Madise. According to Tavast, the impact of malicious content can be reduced by providing developers with diverse and high-quality Estonian-language material. He stressed that creating Estonian-language language models familiar with our cultural context is a shared responsibility. “Only Estonians can ensure that Estonian language and culture do not fall behind in the development of artificial intelligence. We must act today — tomorrow is too late,” Tavast explained.
The creation of high-quality Estonian-language AI solutions depends largely on access to language data. Today, such access is hindered in several fields by legal uncertainty regarding the use of language data, particularly concerning personal data and copyright. Although the legal norms exist, uniform practice and guidance are lacking, which likely prevents data holders from applying the exceptions provided by law. “Our biggest challenge is legal clarity — clarity that helps define where private interest ends and public interest in the survival of the Estonian language begins,” Tavast said.
Chancellor of Justice Madise confirmed that it is the state’s responsibility to create a framework that fosters innovation while protecting people’s rights. “Clear and well-functioning Estonian language is a shared value. To ensure its preservation and development in the era of artificial intelligence, we need a well-considered legal framework,” Madise emphasised. She added that the framework must give data owners the confidence that sharing their valuable linguistic data is safe. “Important developments that ensure the vitality of the language must not stall due to fear of the new and unknown,” she added.
Photos: Aleksander Gužov
