Assessing
the AI Adoption. The Global AI Index Dataset Used to Build, Train and Test a
Machine-learning Algorithm.
Manuela
EPURE1,2,
1 The Academy of Romanian Scientists, Address: 3 Ilfov
Street, sector 5, 050045 Bucharest, Romania; ORCID https://orcid.org/0000-0002-1405-0389 ; mepure.mk@spiruharet.ro
2 Spiru Haret University, Address: 13, Ion Ghica Street, sector
3, Bucharest, Romania
(corresponding author)
Received: October 7, 2024
Revised: October 25, 2024
Accepted: October 31, 2024
Published: December 16, 2024
Abstract: The
paper aims to analyse the AI adoption at the company/country level and the
efforts made to achieve this objective. The necessary changes for the use of
AI solutions involve not only a significant financial effort, but also
attracting talent, building adequate infrastructure, getting governmental
support and, above all, consistent
investments in research and development at the company/country level. The
paper presents the key elements of the measurement process used in
calculating the Global AI Index, as well as the results for 62 countries,
having as an original contribution the creation, training and testing of a
machine learning algorithm, aiming to extrapolate the AI Global Index. Also,
the purpose of the paper is to demonstrate that AI machine-learning models
can be created, trained and tested to achieve a higher accuracy of
forecasting and can be used regularly in the decision-making process. The
scientific journey was possible due to open access to the data used to
determine the AI Global Index, as well as to the use of collective experience
and wisdom (e.g. Google Colab and Python
programming language). Even though the results have just a demonstrative
value encourages the research expansion to calculate the Global AI Index for
Romania, a country which is not listed among the 62 countries for which the
Global AI Index was calculated in 2023. |
Keywords:
artificial intelligence, Global AI Index, machine learning, algorithm, |
Epure, M. (2024) The Global AI Index Dataset Used to Build,
Train and Test a Machine-learning Algorithm. Journal of Knowledge Dynamics, Vol.
1, No. 2, p33-53. https://doi.org/10.56082/jkd.2024.2.33 ISSN ONLINE 3061-2640