Article Preview
TopIntroduction
The general purpose of this scientific paper (Paper) is to analyse the impact of artificial intelligence (AI) and digitalisation on the macroeconomics of the EU member states (including Romania).
To reach the general objective of this Paper according to the research and writing methodology we have chosen, we have selected 6 specific objectives, all deriving from the main one: the analysis of the birth and evolution of the concept of artificial intelligence; identification and analysis of various forms of artificial intelligence, of its possible impact on the aggregate economy, based on the findings of the scientific writings on this subject; identification of the relevant sources of statistic data and information enabling us to undertake an inter-state comparative study regarding AI and digitalisation; analysis of the intensity and sense of the correlation between various indicators expressing the use of AI systems and digitalisation, in the EU members states (including Romania); analysis of the interdependence between the average GDP per capita and labour productivity on the one hand, and the AI and digitalisation levels, on the other; analysis of the interdependence between the share of IT professionals in total employment and the levels of AI and digitalisation.
The Paper is structured in 2 sections, based on research and analyses pursuing to attain the general objective and the specific objectives. They are preceded by an introduction, and followed by conclusions and bibliographical references.
Section 1, titled “Impact of artificial intelligence (AI) on macroeconomics, and how much we know about it” is intended to cast a glance into the genesis and evolution of the concept of artificial intelligence, starting from the definitions given to artificial intelligence in its various forms of action, and to its possible effects on aggregate economies, as identified in a large variety of dedicated literature, cited in acknowledged international data bases, such as JStor, ProQuest, Science Direct, and collections of working papers, among which NBER, REPEC, SSRN).
The results of the research will then be used to develop Section 2: An analysis of the macroeconomic impact of artificial intelligence. Chapter II begins with a description of the general objective of the Paper, the specific objectives, and hypotheses of the research, and elaborates on the techniques used to select and collect information, on the indicators employed for the analysis, as well as a summary of the research methodology.
The initial data basis contains the values of 12 indicators for the year 2020, in respect of 25 member states of the European Union (including Romania), divided into 3 categories: 6 indicators reflecting the use of AI systems in companies operating in EU member states (more exactly, a) the share of companies using the internal analysis of Big Data with the help of machine learning, namely b) internal analysis of Big Data by way of processing natural language, or generating natural language, or recognising the discourse; c) use of robots for delivery of services; d) provision of services with the aid of a chatbot or a virtual agent responding to clients; e) no artificial intelligence system in use, or f) use of at last one AI syste.
Four (4) digitalisation indicators, which are present in the structure of the Digital Economy and Society Index (DESI) (population with digital literacy above the average; IT professionals in employment; corporate entities using software for conveyance of information; and entities that have acquired at least one Cloud computing service). And two (2) macroeconomic indicators, respectively the average gross domestic product (GDP) per capita, and labour productivity.
Among the range of programmes and methods propounded for use in empirical analyses, we quote: statistical software for Excel – LSTAT, RStudio, Agglomerative Hierarchical Clustering (AHC), the least squares method for the generation of the multifactorial regression model. While the testing of the normal distribution of errors, of homoscedasticity, and error auto-correlation will be performed with the aid of the Shapiro-Wilk, Jarque-Bera, White, and Durbin-Watson tests.
The second part of Chapter II is designed to test and analyse, with the help of the methodologies indicated above, the intensity and the sense of the correlation between the indicators reflecting the extent to which AI systems are used, and the indicators of digitalisation. In the second part, the research continues with the extent to which the GDP, labour productivity, and share of IT professionals among total employment depend on the intensity of the use of AI and digitalisation in society and the economy.