The Turkish Language Association (TDK) is developing a new dictionary expected to contain more than 800,000 words, alongside a large language model designed to train artificial intelligence to better understand and “think in” Turkish, TDK President Osman Mert said.
Mert said the projects are part of the institution’s efforts to protect and develop Turkish while adapting to changes brought by digitalization and AI.
The existing Turkish Dictionary contains around 132,000 words, while the new dictionary will seek to document vocabulary across literature, science and different cultural fields, Mert said.
The project is being supported by a balanced corpus containing 1 billion words, which will help researchers identify contextual meanings, proverbs, idioms, terminology and different linguistic structures.
Mert said TDK has also been working for around three years on a large language model designed to develop an AI system that “thinks in Turkish.”
“The theoretical part is complete, and we have now moved to the practical stage,” Mert said.
He said the technical infrastructure is largely complete and data entry has begun, with the data being extensively labeled for use both in AI development and linguistic research.
Another major TDK project is the Türkiye Terminology Bank, which currently contains around 310,000 terms, Mert said.
He added that the bank is expected to open to the public and relevant institutions this year, with plans to eventually expand it into a Turkic World Terminology Bank by incorporating terminology from other Turkic countries.
TDK is also developing software to improve the use of Turkish in digital environments.
One program will detect grammatical, spelling and punctuation errors in Turkish texts and offer suggestions for corrections. Another will scan public institutions’ websites around the clock to identify and report language and spelling errors.
A third system will scan online content to identify words that are not included in the Turkish Dictionary and report them to TDK.
Mert said the three software projects are expected to become operational within one year. He added that TDK is expanding educational initiatives aimed at improving the use of Turkish among children, young people and teachers.
The institution has organized diction courses for teachers in Ankara, with plans to expand the program nationwide.
TDK also launched a 120-hour story-writing workshop for university students in Ankara, while a selection of works produced during last year’s program is currently being prepared for publication.
Mert said growing demand for the programs showed the importance of strengthening awareness of accurate Turkish usage, particularly as internet and social media use continues to shape young people’s written and spoken language.
Addressing efforts toward a common Turkic alphabet, Mert said a 34-letter framework alphabet was adopted in Baku in 2024.
He said Kazakhstan and Kyrgyzstan announced the letters they would use at a meeting in Astana on June 15, 2026, while work on the issue continues in Uzbekistan.
“There is no change to the alphabet used by the Republic of Türkiye. There is no need for such a change either,” Mert said.
