Kazakhstan’s media have lately been full of reports about the success of introducing Kazakh into AI. But they appear to be overlooking the matter of whether the country’s state language is ready for such integration. The optimistic public mood on this topic is driven by the official narrative that focuses on high-level achievements, such as the introduction of the first national supercomputer and the AlemGPT language model. However, the actual ‘readiness’ of the Kazakh language for deep AI integration faces significant hurdles that are often overshadowed by these reports. The official narrative’s leitmotif is that the prospects for the Kazakh language in the AI era are defined by digital sovereignty and standardization, ensuring the language evolves from a ‘low-resource’ digital entity into a high-performance computational tool. 

The main problem is not so much the lack of quality digital data in Kazakhstan, but rather the quality of that data itself. To understand the seriousness of the language situation, let’s give one example. Since we in Kazakhstan are now entering the period when the new Constitution has come into effect, let’s turn to a case related to the previous Basic Law of the country.

Here is what was a noteworthy point. The draft of the previous Constitution was put to a national referendum and debated across the country for nearly six months throughout the spring and summer of 1995. That same autumn, parliamentary elections were held under the newly adopted Basic Law. However, the issue of discrepancies between the Kazakh and Russian texts was only officially raised in 2002, seven years after its adoption.

The initiative came from members of the Mazhilis, the lower house of the Kazakh Parliament. In February 2003, MP Serik........

© Eurasia Review