News

Dr. Vassilis Katsouros, Director of the Institute for Language and Speech Processing at the Athena Research Center, was featured on ERT’s Prime program in a discussion with journalist Elias Sakaris on the developments in Artificial Intelligence, large language models, and the challenges faced by the Greek language in the evolving digital landscape.

 

Dr. Katsouros explained that Large Language Models are trained on massive volumes of data, primarily originating from globally dominant languages such as English, Spanish, and Chinese. In contrast, Greek has a significantly smaller amount of available digital content, which directly affects the development of AI systems for the language.

 

As he noted, the training of Greek AI models relied on both Greek and English data, highlighting how knowledge can be transferred from one language to another. He also explained that language models operate through sequences of linguistic patterns and that, when sufficient knowledge is not available in a specific language, systems may either generate inaccurate information or transfer elements from other languages included in their training data.

 

Particular emphasis was also placed on Europe’s role in the development of Artificial Intelligence. Dr. Katsouros stressed that Europe possesses significant linguistic and cultural wealth, which could serve as the foundation for the development of European Large Language Models and related infrastructures.

Referring to the rapid evolution of Artificial Intelligence, he pointed out that the development of such systems is accompanied by increased consumption of computational resources and energy, underlining the need for more efficient and sustainable technological solutions.

 

Finally, he highlighted the importance of technological sovereignty and data governance, emphasizing that both Europe and Greece need to develop their own AI infrastructures and models in order to strengthen their autonomy within the rapidly evolving digital environment.

 

The interview highlighted both the scientific and societal dimensions of Artificial Intelligence, as well as the importance of investing in European and Greek technological capabilities for the future.