Information Technology expert Besfort Gashi has warned of a major risk to the security of Kosovo’s state data, describing the situation as a “National Digital Emergency.”
In an interview for IndeksOnline, Gashi said that “Kosovo’s state documents are potentially being exposed on foreign servers. Public institutions and their employees are likely using ChatGPT, Gemini, DeepSeek and other artificial intelligence platforms for document processing or spelling correction of documents or correspondence, without realizing that this information within the correspondence and documents is being sent abroad.”
According to the expert, the situation is alarming:
“Recently, the company OpenAI (ChatGPT) publicly admitted that it shares user data in certain circumstances, including information on people with mental health problems and suicidal thoughts, which practically means that every text entered into ChatGPT is stored, analyzed, and processed on servers outside of institutional control.
On the other hand, many users in Kosovo, in addition to Chatgpt, also use DeepSeek – a Chinese artificial intelligence.
“All data uploaded to this AI is stored on servers in China, which means that hypothetically, our official and sometimes confidential information is potentially sent to China in the event of using the DeepSeek application.”
The expert points out that Kosovo’s current laws prohibit the distribution of state information to foreign servers, but they are not harmonized with the AI era:
“Administrative Instruction No. 02/2018 requires that all systems that process state data be approved by the Information Society Agency (ASHI). ChatGPT, Gemini, DeepSeek and other AI platforms are not certified for this.
In practical terms, uploading government documents to these platforms is not only dangerous – it is illegal. This constitutes an unauthorized transfer of government data and a violation of applicable law.”
Gashi demands immediate measures:
1. The use of AI for processing state documents is prohibited;
2. A secure infrastructure for AI is created under the control of ASHI;
3. Sanctions are foreseen for any institutional violation.
“This is not a technological debate – it is a matter of national sovereignty. In the age of artificial intelligence, data is the greatest asset we have. The European Union has passed laws that prohibit any state information from being processed by AI. We just need to harmonize the laws according to EU standards,” Gashi concluded.
