Academician Ali Abbasov's speech at the General Assembly of the Azerbaijan National Academy of Sciences
28.02.2025 /
Dear participants of the General Assembly of ANAS,
We have gathered today together with our guests to discuss the report on the research works carried out at ANAS in 2025. For the second year after the signing of the Decree of the President of the Republic of Azerbaijan Mr. Ilham Aliyev dated July 28, 2022 “On some measures related to the improvement of management in the field of science and education”, the discussion of the second annual report of ANAS is held with the active participation of the Ministry of Science and Education of the Republic of Azerbaijan. This constitutes a necessary format within one of the main points of the decree, “integration of science and education,” and brings together all areas of science. In his detailed report, the Academician Secretary has provided comprehensive information on all areas, including important results obtained. I would like to share with you briefly my thoughts on artificial intelligence, its application and future directions of development. Firstly, the Decrees of the President of the Republic of Azerbaijan Mr. Ilham Aliyev in January on approval of the “Concept of Digital Development of the Republic of Azerbaijan” and the information system “National Spatial Data”, clearly show the wide application of artificial intelligence systems in the development of these spheres in our country and the special importance given to this field at the state level. According to the information presented in mass media, the Ministry of Digital Development and Transportation is developing the National Strategy for the development of artificial intelligence in our country, and this document is likely to play the role of a roadmap for the development of artificial intelligence in our country.
At the same time, the report draws attention to the problems of artificial intelligence, as well as to the research and results obtained in this direction by MSE and ANAS institutes. The work on the development and application of AI in Azerbaijan started long ago, and the research on natural language processing (NLP) technologies began in the 1990s, and its rapid development began in our country, as well as in other leading countries, in the 2000s, and our first result was “Dilmanc”, an artificial neural network-based system for translating text and speech from Azerbaijani into English. This system started to be implemented as a natural language processing system based on a natural language model built on the Azerbaijani language, based on national datasets such as 500 million sentences, 2 billion words and phrases, 10,000 hours of speech and with 110 million parameters. However, in 2023, the ChatGPT system based on the Large Language Model developed by the American company Open AI and the subsequent announcement of generative versions surpassed systems based on national language models, including Dilmanc and others, in terms of both theoretical and practical importance, and today systems such as CPT-4o, LLAMA, BERT, DEEPSTEEK, Grok 3 and the upcoming multimodal GPT-5 have begun to be widely used and have gained worldwide adoption. These systems are also widely used in Azerbaijan. However, the main issue today is the creation and wide application in our country of artificial intelligence software with open source code and the corresponding technical support and infrastructure, which is built on national information resources on the basis of Big Language Models, as well as trained on national datasets (Deep Learning) and contains potential and real capabilities of leading foreign LLM systems. As in the leading countries in this direction, we are concerned with the question of how AI will be used in future scientific research and applications, as well as the issue of its positive and negative consequences. This question has particularly interested scientists after it was revealed that two of the three scientists awarded the 2024 Nobel Prize in Chemistry - Google DeepMind directors Demis Hassabis and John Jumper - have successfully solved the problem of predicting the three-dimensional structures of proteins, an important problem in chemistry and biology, using artificial intelligence. So the question is how scientific research should be conducted as the intellectual capabilities of AI approach and eventually surpass those of humans. This question applies to all fields.
The main prerequisite for solving the problem is the establishment of a national center and infrastructure of artificial intelligence in our country with an open architecture and access to current and future development resources and services of the world's leading systems. In this case, collection of local Azerbaijani information resources and their use in the training system will play a fundamental role in solving local specific problems, our dependence will be significantly reduced, and the requirements of national security, commerce, etc. will be taken into account. Today's report provided information on the Electronic Academy system of ANAS, the evaluation of scientific performance using artificial intelligence, the development of a robot and speech recognition and dialog system in cooperation with the National Aviation Academy, which was demonstrated at the exhibition, etc. Particularly, the Journal Park, Scientific Human Resources Database (İREMB), Digital Scientific Information Resource System (REİRS), Speech Database (Speech to Text) and Central Library Archive Digitization System developed in the “Electronic Academy” system can be widely used in the preparation of datasets in the National Center for Artificial Intelligence.
One of the issues attracting attention in the development and widespread application of artificial intelligence systems is cybersecurity systems. Given that artificial intelligence systems can now operate autonomously, this poses new challenges and threats. The problem of their freedom, autonomous development and the possibility of this development getting out of control is becoming more and more relevant, as we have read a lot about in science fiction. Things are now moving towards this. Indeed, as digitalization increases, cybersecurity is considered a major concern. Today, digital resources in the world are growing at 60% annually. All major systems, modern military systems, banking systems are becoming digital. This is also exacerbated by the negative impact of artificial intelligence. Because we already want to fight cyber threats with the help of artificial intelligence. However, in this case, hackers who carry out cyber attacks are taking advantage of this and perfecting it. In such a situation, a serious danger arises. The current state of cybercrime is already very dangerous. Given that AI can make anti-human decisions autonomously and get out of control, one can imagine the consequences of such threats.
In 2023, the global revenue of the information technology industry amounted to $5.3 trillion. The damage caused by cybersecurity is estimated at $8-10 trillion dollars. This even creates an economic paradox. It turns out that digitalization does not justify itself economically. However, we are not talking about economic damage, in general, even future world conflicts and wars are going digital. In such a case, global security comes to the fore. The main problem of national systems here is the software and hardware of artificial intelligence, data sets, information resources and the deep learning principle built on them.
In such a situation, it is important to create the above-mentioned local system and infrastructure of artificial intelligence, taking into account the aspects of reducing national dependence on artificial intelligence technologies in solving future tasks and for technological and innovative development. It should be noted that this approach is not some new revolutionary phenomenon. A number of countries have already started to create such systems. An appeal to the government on this issue may be included among the decisions of the general assembly at the annual reporting meeting held with the participation of the National Academy of Sciences and the Ministry of Science and Education.
It is also important to note that the implementation of large-scale projects becomes more effective and sustainable if it is built on the platform of “public-private partnership”. But if we add another pillar to this widely accepted platform, the realization of innovative projects on the platform of “public-private-academic partnership” will become even more effective and reliable. The creation of the national center we are talking about on such a platform seems more realistic.
Earlier, Academician Isa Habibbeyli and Minister Emin Amrullayev noted the desirability of improving the quality and efficiency of scientific research, and we, scientists, understand that this is an appropriate remark. It is no coincidence that the role of science and education is emphasized on this platform, and academic participation is of great importance in the development of innovative projects.
When discussing the development of science and technology, the active participation of the private sector is necessary. For example, 50-60% of research and development spending in developed Western countries is accounted for by the private sector. For example, in the U.S., Amazon's R&D spending in fiscal year 2023 amounted to approximately $73 billion, which is about twice the R&D spending of the CIS countries combined.
Here are some examples of projects implemented at the Institute of Control Systems on the basis of this platform:
- SCADA system developed in the Institute and Special Design Bureau (self-financing LLC subordinated to the Institute) for monitoring and control of technological processes in “Azneft” PA of SOCAR;
- development of hardware and software of the Unmanned Aerial Vehicle Control System at the National Aerospace Agency;
- Virtual Assistant System operating on the basis of the Azerbaijani language model based on artificial intelligence technologies at the State Examination Center.
These are real examples of application on the basis of economic contracts. Implementation of these works is carried out within the framework of cooperation with SINAM, startup companies Robotroniks, IntelPro and Neorotime. We would like to note that expenses for scientific research on the implementation of the mentioned projects are financed, in addition to the articles stipulated in the state budget, earlier at the expense of ANAS, and now at the expense of Scientific Programs of the Ministry of Science and Education of the Republic of Azerbaijan, grants of the Azerbaijan Science Foundation and SOCAR.
We hope that the upcoming work on the establishment, utilization and development of the National Infrastructure of Artificial Intelligence will be successfully implemented on the basis of the “public-private-academic partnership” platform.
Thank you for your attention!
Academician Ali Abbasov
February 20, 2025
