What is Artificial Intelligence (AI)?

Artificial intelligence (AI) is a technology with human-like problem-solving capabilities. AI in action mimics human intelligence: it can recognize images, write poetry, and make predictions based on data.

Artificial intelligence includes systems and algorithms that can perform tasks that traditionally require human intelligence, such as data analysis, decision making, pattern recognition, and learning.

The role of AI in the modern world.

Modern technologies play a big role in the renewal and modernization of the world. An important place among technologies is occupied by the field of artificial intelligence. Artificial intelligence is the imitation of human intelligence in machines that are programmed to think and act like humans. It involves developing algorithms and computer programs that can perform tasks that initially require human intelligence (visual perception, speech recognition, decision-making, and language translation). Artificial intelligence covers many areas, including machine learning, natural language processing, and robotics.

Currently, artificial intelligence is considered one of the most rapidly emerging and improving areas in science. Artificial intelligence is now used in almost all areas of human activity, enabling automation, improving decision-making, increasing efficiency and productivity, and creating new opportunities for innovation and growth in various industries, including healthcare, finance, manufacturing, transportation, e-commerce, education, and many others.

Artificial intelligence is now widely used in information systems, enriching their capabilities. Machine learning and deep learning algorithms allow systems to recognize images, process natural language, and predict trends based on data. These technologies have revolutionized the way big data is analyzed and used, making information systems more intelligent and adaptive.

One of the main ways that artificial intelligence transforms information systems is through big data analysis. Thanks to machine learning and deep learning algorithms, artificial intelligence is able to process and analyze huge amounts of data, identifying patterns and trends in them that a person may miss. This allows businesses to make more informed strategic decisions and predict market trends. Artificial intelligence enables information systems to create personalized user experiences. AI is widely used in various fields, such as medicine, transportation, financial technology, and, of course, political governance.

The concept of history and the development of artificial intelligence.

Early stages of AI development.

Since the late 1200s, there have been attempts to create an artificial person and his mind. Inventor Raymond Lully designed a machine consisting of circles marked with letters and painted in different colors, which symbolized various concepts, elements of the elements, subjects and objects of knowledge. Various combinations of them were used to derive «formulas of knowledge» using logical operations.

In the 1940s, with the advent of electronic computing machines, artificial intelligence was reborn. Research in the field of artificial intelligence has two goals: to find out the essence of natural intelligence (human intelligence); and to use machine intelligence to transform new knowledge and solve intellectual problems.

In the late 1950s, the first neural networks and neurocomputers began to be developed and created by American scientists U. Maccalock, W. Pitts, and F. Rosenblatt, who represent and currently represent the Neurocomputer Division of the Russian Academy of Sciences.

In 1943, W. MccOluck and W. Pitts proposed a model of a formal logical neuron that could be in two stable states. In 1949, D. Hebb developed a simple rule that allows you to change the weights of connections between neurons in order to train them. In 1951, M. Minsky and D. Edmonds developed a neurocomputer that contained 40 neurons.

The term «artificial intelligence» was proposed at a seminar at Dartsmouth College (USA) in 1956. The first work on AI was carried out at the Massachusetts Institute of Technology under the direction of M. Minsky and J. P. Blavatsky. McCarthy, at Carnegie Mellon University under the direction of G. Simon and A. Newella. They are considered the «fathers» of AI.

Heuristic search and proof of theorems (1956-1969). During this period of AI development, the main research and development activities were presented by J. R. R. Tolkien. McCarthy, J. Robinson, K. Green, D. Hebb, F. Rosenblatt, M. Minsky: development of the Lisp language (J. McCarthy).

Knowledge representation (1969–1979). The DENDRAL program, developed in 1969 by E. Feigenbaum, B. Buchanan, and E. Liederberg, contained detailed information about the field of organic chemistry and helped specialists determine the molecular structure of organic compounds based on data obtained using a mass spectrometer. The PROSPECTOR expert system (1979), used in geological exploration, was a huge success. With the advent of expert systems, business in the field of intelligent information technologies is becoming profitable for the first time. In the PROSPECTOR system, the knowledge base was represented as a semantic network, and the system provided interaction with the user in natural language.

Commercial success of the computer industry 1979–1986.

The first intelligent system to be used in industry was the K1 expert system, developed by McDermott in 1982. The K1 system was used to configure computer systems of the VAX family. The commercial version of the system, developed by Digital Equipment Corporation in collaboration with Carnegie Mellon University (USA), was named XCON. By 1986, this system allowed the corporation to save $ 70 million annually. In addition, the use of the system reduced the number of errors from 30 % to 1 %.

In 1981, Japan announced the start of a project for 5th-generation AI-based machines. This project has helped boost AI research in many countries. Since 1985, expert systems, and then systems that perceive natural language (EJ-systems), and then neural networks (NS) have been actively used in commercial applications.

Stage of development of intelligent systems of the second generation (1996–2000).

Since the early 1990s, AI has been dominated by two major trends: integration and decentralization^[209].

Artificial intelligence in public administration.

Artificial intelligence is used in public administration primarily for the systematization of documents and automation of bureaucratic routine processes. For example, neural networks help optimize search, monitor data on the Internet, and register citizens ' applications.

AI embedded in various programs simplifies interaction with clients, as well as working with content and documents. So, the SaluteBot product SaluteBot, which works together with GigaChat, is able to create chatbots for users to answer various questions. The system can work in any organization, including the state one.

How is artificial intelligence used in public administration?

The amount of data that government agencies work with is growing rapidly, and AI can improve the efficiency of their work. For example, improve the quality of service, which increases their confidence in government agencies.

Artificial intelligence technologies in public administration contribute to increasing transparency and accountability of public authorities. Automation of processes and the use of analytical tools allow you to more accurately track the progress of tasks, identify problems and solve them.

Examples of artificial intelligence for public administration. Today, the articles cite reports from experts on the introduction of AI in Russian government agencies. So, recently it became known that in the Sakhalin region of the Russian Federation, a system with AI for master planning will be launched, the purpose of which will be to speed up procedures for the development of territories.

The use of artificial intelligence in public administration is not limited to reducing bureaucracy, it even helps to fight crime. For example, in the UK, the police use the AI for Social Good system based on AI, which collects information about crimes, helps to predict and prevent them.

Artificial intelligence is also used for traffic management. In Barcelona, the Smart Traffic Lights system uses data from cameras and sensors to optimize the operation of traffic lights, which helps reduce traffic jams and improve transport infrastructure.

GigaChat is a neural network model for creating content, generating and summarizing text, writing code, maintaining a dialog, classifying data, and intelligently processing information. With the GigaChat API, you can optimize your business processes, speed up your company's systems, and expand their functionality.

Programs and applications with AI in public administration in foreign countries.

The list of options for using artificial intelligence in public administration is rapidly expanding.

In Russia, in 2024, an AI-based digital assistant was launched on the «Gosuslugi» website, which the developers called «Max». The tasks of this «assistant» include answering questions from citizens, talking about benefits, pensions, and fines. Such programs operate all over the world.

In Estonia, the e-Tax system uses AI to accept online and verify tax returns from citizens.

In China, Smart Grid networks use AI to analyze data on energy consumption and automatically adjust its supply to facilities.

In Japan, Smart Waste Management systems help to sort waste and improve the environmental situation in cities.

In California, the AI-powered FireMap tool analyzes weather and vegetation data to predict fires.

In the United States, the RPA (Robotic Process Automation) neural system is used to automate the verification of tax returns, which reduces the processing time of applications.

Singapore has a state-owned chatbot, Ask Jamie, which answers citizens ' questions and helps them find the right information. This has reduced the waiting time for answers. It uses natural language processing (NLP) to understand and answer citizens 'questions.

These examples demonstrate how AI can improve the efficiency and quality of public administration, making it more transparent and citizen-oriented.

The future of public administration with AI.

Artificial intelligence will play a key role in the development and implementation of new directions. Mathematical models built on the principle of neural connections will inform drivers and the state traffic Inspectorate about traffic flows and other aspects of urban life. This will improve the work of city services and the quality of life of citizens, and reduce the cost of city management.

AI systems will be able to analyze data on natural disasters, epidemics, and other crises to predict their occurrence and suggest measures to prevent them. This will allow state emergency services to take measures to protect the population in advance.

The structure and functions of government agencies may also change, as many routine tasks will be automated and the need for administrative staff will be reduced. At the same time, you will need specialists in data analysis, cybersecurity, and AI management. This, in turn, will require retraining of employees and changes in the organizational structure of state bodies.

The introduction of artificial intelligence in public administration helps to develop innovations and increase the country's competitiveness in the international arena. States that actively use AI in management can adapt to new conditions faster, develop and implement advanced technologies, accelerating the country's economic growth and development.

AI technologies are actively developing. The so-called generative AI, which is used to create images (neural networks like MidJorney), texts (ChatGPT, YandexGPT), videos (Pictory neural network, etc.), as well as intelligent decision support systems (solutions for processing data for a given goal), is particularly successful and popular today. Given such extensive functionality, these technologies are increasingly used in various spheres of human life, including politics.

Using AI in politics.

Here are some examples of the use of AI technologies in politics.

President Barack Obama's 2008 presidential campaign in the United States was the first campaign to actively use social media. During his second presidential campaign in 2012, AI technologies were already used to calculate the best day, state, and audience for a public speech by Barack Obama. According to various estimates, this provided an advantage of 10-12% of the vote.

In 2016, the British company «Cambridge Analytica» illegally collected data from 87 million Facebook users for analysis, working closely with the headquarters of Donald Trump.

According to a study by the University of Oxford, in 2020, digital technologies were used to manipulate public opinion and spread misleading propaganda in at least 81 countries. The impact tools included chatbots, microtargeting, content generation algorithms, cloned human voices, and databases for facial recognition.

For the most part, Russian politicians are just beginning to explore the possibilities of AI technologies. The LDPR presented the Zhirinovsky neural network, which imitates the words and speeches of the former party leader. United Russia is considering the possibilities of neural networks for analytics, forecasting, and generating video materials and images. At the same time, Deputy Secretary of the General Council Sergey Perminov places special emphasis on threats and fakes. The Communist Party of the Russian Federation is studying the possibilities of AI for use in propaganda work. «Fair Russians» were highlighted in the media with the topic of a possible ban on the domestic neural network «Kandinsky», and «New People», on the contrary, defended AI as a technology of the future.

The main directions of using AI for political tasks.

AI technologies can be implemented at almost all stages of political processes. The main areas of application of AI technologies include:

Create campaign content (texts, images, videos). AI can analyze the information field and formulate various slogans and campaign materials that are relevant to the current situation. This can be a speech, a press release, a photo, or a video. AI can also be used in the context of a developing communication crisis, where it is necessary to track news, identify patterns and risks to minimize damage. The response time may be reduced to minutes, not hours or days. But so far, the generated materials do not take into account the socio-cultural context and are more like the work of a student or intern that requires editing.

Sending out targeted messages. In the future, texts that were personalized based on the user's search history and preferences will become more detailed. For example, in American political campaigns, targeted messages are used to influence clearly identified swing groups of voters. This allows candidates not to spend money on undecided voters, but to focus on undecided voters who will decide the outcome of the election.

Use chatbots to answer questions from voters. A trained chatbot will be able to save time for phone operators and answer the most frequent questions from voters. To effectively persuade different groups of voters, the complexity of responses can range from emotional appeals and slogans to statistics, law enforcement experience, and global experience. In addition, representatives of political parties will be able to «collect» their voters on special digital platforms and communicate with them. Behavior analysis will also be conducted there (with the consent of the voters themselves).

Modeling of political preferences.

AI is able to analyze large amounts of data to shape the political agenda. For example, in Denmark, the Leader Lars artificial intelligence analyzed the publications of 230 small political parties since 1970, and created a Synthetic Party program based on this data, the agenda of which reflects the political preferences of about 15-20% of the country's voters. Thus, AI can highlight realistic requests from voters on the political agenda, forcing live politicians to pay attention to them.

Improving predictive models. AI can help predict the results of elections and other political campaigns, complementing already known predictive technologies by highlighting non-obvious human interdependencies.

Conducting opinion polls. It is possible to use AI robots for telephone surveys of the population about political preferences. However, so far, robots do not recognize human speech well enough, do not respond well to changes in the respondent's behavior, and make many mistakes.

Analysis of fakes. Checking the validity of content can become one of the most popular areas.

This list is not exhaustive. Any task related to the analysis of a large amount of data and routine procedures can be given to AI in the future – whether it is analyzing the results of voting, working with an array of legal documents (checking signatures, checking documents submitted by candidates, etc.). However, the substantive core of the profession of political consultants – strategy and ideology – due to its complexity and creative nature is now beyond the control of AI AI.

AI risks and regulation.

Large-scale AI penetration will be seriously monitored by regulators, as the use of these technologies carries too sensitive and poorly predicted risks. This process will be accompanied by a discussion about the ethics of artificial intelligence.

If Athenian democracy could be carried out «face to face», today political communications have reached such a degree of virtualization that sometimes we cannot verify the source and accuracy of the information disseminated.

Neural networks can create hyper-realistic human voices, images, videos, and audio in seconds at minimal cost. Linked to powerful social media algorithms, targeted emails, texts, or videos can be used to mislead voters on a scale and at a speed that was previously impossible. You can only imagine what it might be: a call from a candidate of the race advising you to vote on the wrong day; the appearance of incriminating videos or audio recordings.

Researchers from Stanford University conducted an experiment to determine the credibility of AI. For the experiment, the neural network created texts on several controversial topics (such as supporting the arms trade or banning abortions). Then these texts were mixed with real ones and given randomly to participants of the experiment from different categories of the population. Participants were asked to state their position on the questions before and after the reading. In all the comparisons, the messages generated by the AI were convincing to readers. Moreover, the subjects noted that after reading them, they began to support the position that the AI defended more.

Most importantly, AI services provide sophisticated tools for the average person. People no longer need to be programming experts or video masters to generate a digital product. They don't have to work on a «troll farm» to initiate large-scale political debates on the Internet. They can simply use advanced technology to spread the messages they want (including fake ones). In this sense, anyone can become a creator of political content and try to influence voters or the media.

To address these shortcomings, government regulation is required. It should be based on certain ethical principles, such as accountability, transparency and fairness – what is now called «responsible AI» in the world practice.

To create an AI regulatory model, you need to answer the following questions:

- how consent to the use of users personal data is obtained and confirmed;

- how user data is depersonalized during AI training and operation.

- who can access the technology (all or selected).

- to what extent censorship is appropriate, and who will provide it (the state, developers, or special organizations).

It is worth noting that the answers to these questions may vary from country to country. That is why states can regulate AI activities in different ways.

The US Department of Commerce has initiated a public discussion on the need to certify potentially risky new AI models before they are released. The G7 countries announced that they are working on international standards for the use of AI. The EU is also developing its own document to ensure that «AI developed and used in Europe is fully compliant with EU rights and values, including human oversight, security and privacy».

On May 16, Sam Altman, CEO of OpenAI, the company that created ChatGPT, called on the US Congress to prevent AI from causing «significant harm to the world» and «manipulating» the US presidential election. At the same time, US lawmakers began to consider a rule requiring labeling of pre-election ads created using AI. It is noteworthy that in this regard, the deputies of the State Duma of Russia went the farthest, who proposed labeling all content created using neural networks.

If Western countries are inclined to more subtle regulatory regulation, then China is inclined to censorship. At a time when the US authorities are more concerned about the fact that neural networks give deliberately false answers, China introduces censorship in neural networks, rather than regulating them. The Cyberspace Administration of the People's Republic of China has published a draft of measures to regulate generative AI services. The content created by them should reflect the basic values of socialism and should not call for undermining state power or overthrowing the socialist system.

In Russia, the discussion about limiting the use of AI in sensitive areas has been going on for several years. Therefore, in August 2021, the United Russia party adopted the Digital Manifesto, which outlined the risks of using AI (mainly in such areas as education, medicine, law, property, and the right to work).

In general, Russia seeks to combine both Western approaches in its desire to create certain regulatory norms for AI (an AI Code of Ethics has been developed, and industry standards in the field of AI are being implemented), and Chinese practices.

The impact of China's experience was once again confirmed at SPIEF by the head of the Ministry of Digital Development Maksut Shadaev, who said that Russia's approaches to regulating artificial intelligence are close to Beijing's position, where they believe that all data belongs to the state and is a strategic resource.

---

^[209] Romanov R. Artificial Intelligence in the Process of Making Foreign Policy Decisions // RSMD. 29.07.2022.- URL: https://russiancouncil.ru/analytics-and-comments/columns/cybercolumn/iskusstvennyy-intellekt-v-protsesse-prinyatiya-vneshnepoliticheskikh-resheniy/. Date of access 26.03.2025