=Paper=
{{Paper
|id=Vol-3698/paper4
|storemode=property
|title=Human Intelligence vs Artificial Intelligence: Secondary Term Creation in Information and Communication Technology Field
|pdfUrl=https://ceur-ws.org/Vol-3698/paper4.pdf
|volume=Vol-3698
|authors=Dace Šostaka,Juris Borzovs,Jānis Zuters,Oksana Nikiforova,Vitaly Zabiniako,Arnis Staško,Jānis Grundspeņķis,Ali Mert Erdoğan,Ourania Areta Hiziroglu,Abdulkadir Hiziroglu,Anup Bera,Sujaya Kundu
|dblpUrl=https://dblp.org/rec/conf/balt/SostakaBZ24
}}
==Human Intelligence vs Artificial Intelligence: Secondary Term Creation in Information and Communication Technology Field==
https://ceur-ws.org/Vol-3698/paper4.pdf
Human Intelligence vs Artificial Intelligence: Secondary
Term Creation in Information and Communication
Technology Field ⋆
Dace Šostaka*,†, Juris Borzovs† and Jānis Zuters†
University of Latvia, Faculty of Computing, Raina bulvaris 19, Riga, LV-1586, Latvia
Abstract
Information and Communication Technology (ICT) terms are predominantly created in source
language (English) and then secondary–created in other languages (Latvian, Lithuanian and
others).
In first part we focus on insight into ICT secondary term creation (STC) process in Latvia in
general and analyses of the process in particular, research done of the term creation in general
and STC in ICT field in particular, and introduce to the possible steps of STC process to be
optimized.
The second part is devoted to case study: ISO standard “Artificial intelligence concepts and
terminology ISO ISO/IEC 22989:2022” in general and “dependable and dependability” in
particular. We contrast and compare human approach to term creation vs the current
possibilities of applying existing artificial intelligence (AI) tools to term creation.
In conclusion, we outline the most productive ways for automating parts of the STC process of
ICT in Latvian and provide insight into possibilities for further development.
Keywords
Artificial intelligence, Information and Communication Technology, secondary term creation,
Latvian, English1
1. Introduction
By way of introduction, it can be said that if in a way the Three Laws of Robotics [7],
applicable in the broadest sense of the word to what is called artificial intelligence (AI),
were already described in the 1960s and 1970s by Isaac Asimov in his “I, robot” and
Arthur C. Clarke in “2001: A Space Odyssey” [8] by personification of HAL 9000; if B.C. and
A.D. stand for Before Christ and Anno Domini, respectively – a beginning of a new,
different era then similarly, within the last few years, since the access to the open AI
Baltic DB&IS Conference Forum and Doctoral Consortium 2024
∗ Corresponding author.
† These authors contributed equally.
dace.sostaka@lu.lv (D. Šostaka); juris.borzovs@lu.lv (J. Borzovs); janis.zuters@lu.lv (J. Zuters)
0000-0003-2066-937X (D. Šostaka); 0000−0001−7009−6384 (J. Borzovs), 0000-0002-3194-9142
(J. Zuters)
© 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
33
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
technologies has been provided, in a way it can be considered that a new time counting
system has been created within regard to gathering information, processing it and
creating a new information: namely, before the release of various types and subtypes of AI
and after the release.
General futurological speculations about the pros and cons of AI systems in human life
in general and language in particular [14] in the broadest understanding of the concept
were especially popular in the 1960s and 1970s, predominantly by science fiction writers
[13], a detailed research article appeared already in 1980s, dealing with applications of AI
in terminology work [12], describing the steps to be taken with a rather precise
prediction.
When researching the development of Information and Communication Technology
(further in the text – ICT) terms, the main aims and objectives of this article are as follows.
First of all, provide a brief insight into the research in the term creation in the scientific
community. Second, describe the process of secondary term creation in Latvian in general,
in the ICT field. Third, the detailed process of term formation is described, looking for
aspects that could be automated. Fourth, in the case study, the methods for parts to be
automated were researched, compared and contrasted: efficiency of writing tailored
programs vs using already existing AI tools. Fifth, we draw preliminary conclusions about
the research and look at possible future research avenues.
1.1. State of Art: Automated Terminology Creation
While looking for relevant research regarding the automation of terminology, we looked
into the latest publications of the leading academic journal “Terminology” and even if
there could be found articles devoted to the automatic term recognition [32] and
extraction [33]; also there is research, dedicated to the possibilities of automating the
process of creating dictionaries [34]. There is also pertinent research on
the possibilities of automated term creation [35] and of secondary term creation (also -
localization) of medical terminology in the German language [36]. Nevertheless, there
can’t be found published research on the possibilities of automation of secondary term
creation.
1.2. Research in AI and Terminology
Within the last few years, there has been an understandable explosion of specific research
regarding the application of various AIs on a general, popular science level [11]. Specific
research [24], published in 2024, is already devoted to linguistics and AI and to aiding AI
in terminology in particular [25].
On the one hand, in the last decade, numerous researches have been carried out on ICT
terms from various perspectives and in different languages: in 2014, Albanian on
codification of Computer Terminology [3], Romanian on linguistic approaches when
creating IT vocabulary [4], and Serbian on secondary term creation in 2016 [6].
34
Although there is research regarding the term formation per se in various languages
available, the number of research narrows down with the specialisation of the field in
secondary term creation for ICT terminology.
There is research in the 2020 on state of art, regarding Europe and languages [18]: in
the article about the leaders of the European language technology in 2020, In the article it
is emphasized that following leading trends can be observed: already developed language
technologies for Latvian, Lithuanian, Estonian and other languages promotes visibility for
languages on digital landscape and facilitate their sustainability and development.
1.3. Research in ICT Terminology and its Creation
There is available bilingual research about the translation of IT terms in Swedish [20],
where the Swedish experience of forming English ICT terms into Swedish described - as
the main method is outlined borrowing and loan translation. In Romanian, there has been
carried out research focusing on hybrid terms in computer science [21], regularities and
norms in secondary term formation [22] and phraseology in Romanian ICT texts [23].
Also, there is research from the Polish language perspective, devoted in general to
linguistics and AI [24] and terminology work [26], terminology work and possibilities of
generative AI for low-resource languages [25].
There has been carried out research on English computing terminology as a system, in
2010, as well as research on trilingual term creation in ICT field of correspondingly
English, French and Igbo [27] and English, Russian and Kazakh [28]. There is also available
research on the Computer science (ICT terms) in academic discourse, analysing 444 most
often used words in Computer science publications (2023) [19].
Our previous research focused on Latvian ICT terminology creation principles and
observation of the implementation of ICT terms in general [29] and, in particular, to a case
study of Computer-Assisted Latvian ICT terminology development [30].
It is worth emphasizing that even though there are terms that have multiple meanings
in the source language (English) we secondary recreate the concept. We work with ISO
standards (ISO/IEC/IEEE 24765:2017(E), “Systems and software engineering –
Vocabulary” and ISO/IEC 22989:2022(en) “Information technology – Artificial intelligence
– Artificial intelligence concepts and terminology – ISO standard”).
2. Case study: Human Intelligence vs Artificial Intelligence
In order to understand the actual process of the term creation and distinguish the places
to be optimized, the following avenue of research was chosen in the following four steps.
Firstly, an analysis of the terms recently discussed and accepted in our commission was
run through theAI tool (ChatGPT), comparing the answers provided by AI and the decision
of the terminology Commission.
The percentage of correspondence was checked and creativity aspect for the terms that
did not return identical answer with our commission’s decision was checked.
Two ways were chosen. Prompting (phrasing of the question to the AI system) was
used in order to devise a meaning for the cases when both the meaning of the term and
35
form of it was wrongly represented and B, a program for automated extraction of the
terms from AkadTerm (Academic term database) was created for checking the results
Secondly, a detailed transcript [10] and analyses of the term discussion and accepting
from actual term commission meeting was carried out, looking for sources consulted (that
could be automated), for extended clarification moments of the meaning of the term and
the concept it denotes.
Then AI tools were used in checking the possible solutions for the dependable and
dependably and checking the obtained results for validity. Full prompting history is
available online [9].
First, we explored the actual term formation process in Latvian ICT, analyzing the most
typical scenarios. Determined the actions (steps) of the process that could be automated
and distinguished two main parts of STC: technical aspect (searching for existing
equivalent/s of the term) and creative aspect (looking for and/or creating new term) and
focused upon optimizing the search for existing terms and/or term parts and case study:
pilot project looking for ways of optimizing the search for existing terms and/or term
parts.
The most often scenario of STC is searching in AkadTerm (Academic Terms database)
[16] as a first resource, looking if term or part of term as been accepted by the Information
and Communication Technology Terminology Subcommission, the Latvian Academy of
Sciences (further in the text – the Commission). The most often used method for STC is
combining already accepted terms (if they reflect the meaning of the concept) in a new
term.
2.1. Computer Program for Extracting Terms from the Academic Terms Database
When manually searching for the terms, the process looked as follows, when preparing for
the Commission meeting. Term preparation consists of repetitive steps, four of them are
mainly mechanical:
first, copy the source term, paste it into search field (of AkadTerm),
second, check returned results a)if the term has already been accepted by the Commission
and b) check if the new term is part of already accepted terms.
third, if the term cannot be found as accepted, then the IATE, Microsoft Terminology and
other sources are searched.
fourth, copy the results in the end document and continue the process in other sources.
Pilot project was created, comparing the time needed for manual work and with the aid
of a computer program. For the manual implementation of the four abovementioned steps,
when preparing 40 terms, the time invested was 4 hours.
The computer program for testing the idea was created, largely imitating the four steps
taken manually, when looking for ways of reducing time for the term extraction and
combining the returned results:
First, “search and return Commission’s accepted terms”:
Second, “search and return ALL terms, available via AkadTerm: Commission, Microsoft
Terminology and others”
Third, “search and return: compound terms; partial matches.”
Forth, “search and return: compound terms; full matches.”
36
Summary. For preparing 40 terms with aid of program, the time invested was 1 hour.
The preliminary results were presented in the 2023, in 5th international conference on
terminology “Scientific, administrative and educational dimensions of terminology” [15].
The first results were promising and we decided to try out AI approach in facilitating the
secondary term creation. The process will be reflected in the next subchapters.
2.2. Dependable and Dependability
In order to understand in detail the exact process of not-so-trivial case of STC, we analysed
the discussion of two terms that have posed a challenge for some time because of difficulties
in reaching consensus. Thus, a more detailed analyses of the discussion process is reflected.
When discussing the dependable and dependability (ISO/IEC/IEEE 24765:2017(E),
Systems and software engineering – Vocabulary), first of all the definition and context is
given for understanding of the terms.
2.2.1. Determining the Term Context
In the case of the terms, the main context was provided by commission member [9] and it
was explained that “historically, «reliability» was created as a concept long ago. There has
been already discussions regarding offering a completely different word, with another root
and term already in use, «drošums» (safety, security). Nevertheless in Latvian «drošums»
gravitate in its meaning very much towards either English «safety» or something related to
«security», so it is not a good choice for «dependability» and «dependable»” and the
commission shall look for different word.
Further, it was explained that “if the meaning of the English word «reliable» is checked,
then «uzticams» (trustworthy) is returned in Latvian. So, it shall be understood what this
term means on the conceptual level in this field, namely, in the field of systems. Looking
further into the meaning of the term, the answer is as follows: «It is the ability of a system to
perform/implement defined functionality within a defined timeframe», and it is usually
expressed as a percentage of the total time the system exists. It is a long-established method,
and it is nothing [conceptually] new.”
A chronological approach to the understanding of the terms is taken, and it is explained
that “In its turn, «dependability» came relatively later from various sources, primarily
standards; I don't even know all of them. There is even more confusion and chaos with
«dependability» because, in some sources, it is interpreted as an utterly synonymous term
with «reliability». But mostly, there are some other attributes.”
The etymology of the terms is briefly discussed and the documents they originated are
emphasized, one coming from a political background and another from technical
documentation “[..] I assume «dependability» might be more like a political term. It could be
originating from different political documents. «Reliability», on the other hand, has been
strictly a technical term from the very beginning.”
2.2.2. Checking the Meaning of Latvian Terms
Then, the problem with searching for an appropriate term in the target language is defined
as following: both terms have the same translation into Latvian, namely “drošums”: “Thus,
37
the problem is very clear-cut: if we take a general language dictionary and check both of
these terms, namely «dependability» and «reliability», then we find as a translation the same
Latvian word «drošums»; it is unacceptable because it creates confusion in understanding
and that is the problem.”
The actual reasoning for the need of clear term is clearly defined in the very beginning
of the discussion as “[..] there should be a new term for «dependability» because it is a more
recent term that appeared later. Something very similar, but not the same.
And there we have the linguistic problem: we haven't found the word for the Latvian
equivalent of «dependability».”
2.2.3. Consulting External Sources
The next step is a discussion among the terminology commission members about the
individual understanding of the terms «dependability» and «reliability». It shall be noted
that this is the most time-consuming part of the discussion where various approaches
while looking for a clear understanding of the term, are used: distinction of the terms,
aiming to determine dominating aspects “distinction is to be made between «dependability»
and «reliability», where the former includes the latter as one of its aspects, then I would
rather say that «dependability» as a broader term should be translated as «uzticamība»
(reliability) and «reliability» as «noturība» (stability, resilience).”
Then the offered term “noturība” is back translated from Latvian into English,
emphasizing that “«noturība» meant something else and it wouldn't reflect the concept in
this case.”
The possible translations of the “noturība” in English are searched for and another
commission member provides the answer that “Latvian «noturība» in English can be
translated as «persistence», «resilience»; «stability», and «immunity».” [9]
The same process – offering words that could reflex the meaning of the term in Latvian
and the backtranslation of them into possible English equivalents - is provided for other
Latvian words, offered for discussion: noturība, uzticība, paļāvība, drošums, pārliecinošs,
vērtums, vērtība.
2.2.4. Creating Compound Terms
Then, when it is not possible to find one word that would reflect the concept, the next step
is taken, creating the compound terms: vispāruzticams, vispāruzticamība, paļāvībvērts.
It shall be mentioned that the offered words and compound words are checked in the
language system: how they can be inflected and how they can be used in the sentences. It
should be possible for a term to be used in sentences; thus the adjective form is checked.
In the course of the discussion various sources are checked, in order to clarify the meaning
and justify the offer: dictionaries, term bases, also AI system, looking for terms in different
contexts. The decision commission members agreed upon, are «dependable» as
«paļautiesvērts», «dependability» as «paļautiesvērtums».
To sum it up, it can be seen that there are a lot of searching, comparing and contrasting,
confirming the provided ideas, looking for context. Thus the idea for AI fine-tuned to
compare the results from AI and our accepted terms.
38
2.3. ISO Standard ISO/IEC 22989:202
For research purposes we choose an ISO/IEC 22989:2022(en) “Information technology –
Artificial intelligence – Artificial intelligence concepts and terminology – ISO standard” that
we have been discussing in our commission from the October, 2023. The total number of
the terms in the ISO standard are 140. The further results in this subchapter were first
represented in Baltic Digital Humanities Forum April 25-26, 2024 [31].
The experiment was carried out as follows. First we ran through all the terms that are
in the ISO standard with definitions; looking from answers in AI system ChatGPT4.
For each English term (e.g. “genetic algorithm”), with its definition (“algorithm which
simulates natural selection by creating and evolving a population of individuals
(solutions) for optimization problems”):
1) ChatGPT was prompted to carry out secondary term formation (e.g. “Provide a Latvian
translation for an ICT term “genetic algorithm”, with following definition (see above) and
the result was recorded (“ģenētiskais algoritms”);
2) “manual” secondary term formation by the Commission was carried out (e.g.
“ģenētiskais algoritms”);
3) results of the two processes were compared and categorized as exact match (Table 1)
with 75 terms, partial match (Table 2) with 65 terms and no match (Table 3) with 5 terms.
Table 1
Both concept and grammatical forms are correct.
No. Term (ISO, English) Term (ChatGPT, Latvian) Term (Commission, Latvian)
49. data quality checking datu kvalitātes pārbaude datu kvalitātes pārbaude
[…] […] […]
66. decision tree lēmumu koks lēmumu koks
[…] […] […]
24. genetic algorithm ģenētiskais algoritms ģenētiskais algoritms
[…] […] […]
Table 2
The concept is formed almost correctly. There are stylistic differences or/and minor
grammatical mistakes.
No. Term (ISO, English) Term (ChatGPT, Latvian) Term (Commission, Latvian)
65. Bayesian network Bejas tīkls Beijesa tīkls
[…] […] […]
48. data annotation datu anotācija datu anotēšana
[…] […] […]
39. procedural knowledge procedurālas zināšanas procedurālās zināšanas
[…] […] […]
39
Table 3
The concept is formed incorrectly, even if a part of the term is translated correctly.
No. Term (ISO, English) Term (ChatGPT, Latvian) Term (Commission, Latvian)
8. application specific konkrētai lietošanai lietojumam pielāgota
integrated circuit pielāgots integrētais integrētā shēma
shēmviens
[…] […] […]
13. cognitive computing kognitīvā datortehnika kognitīvā datošana
[…] […] […]
54. ground truth pamattiesība mērķa vērtība
[…] […] […]
It can be concluded that:
1. Exact match usually refers to the verbatim (word-for-word) translated terms or terms
accepted some time ago. Partial match requires post-editing, still it speeds up the term
discussion process. No match requires manual research in the term databases and other
sources.
2. No significant element of innovation has been detected so far.
3. Nevertheless, AI can find and mechanically combine terms found in its sources, thus
facilitating the groundwork for discussing terms and freeing up more human resources for
the creative part of secondary term formation.
AI tools can be used as a tool for the first checking of the new material and then, the
remaining part of the terminology units are checked with the computer program from
reliable sources instead of attempting to obtain all results from the AI. Thus, hybrid
approach, combining the best from both worlds, works the best in order to speed up the
term creation process. Generalized conclusions can be at the ending of the article.
2.4. AI Approach: Dependability and Dependable
As a concluding phase of the current research process, in order to find out what manual
tasks can be delegated to the AI when preparing terms for discussion and during the
discussion, we asked the MI system prompts that were, as far as possible, similar to the
questions and clarifications that were made during the Commission meeting.
Of course, it was not possible to completely duplicate the discussion during the
terminology commission meeting and the prompts to the AI system but we summarised
the main points:
1) clarifying the conceptual understanding of the terms in source language (English),
2) searching for possible duplicating terms in source language, corresponding to the
definition provided in the source language,
3) translating in the obtained terms in target language (Latvian),
4) looking in obtained results in the target language (Latvian) the possible useful
propositions for the terms that could be used to reflect the meaning.
40
2.4.1. Definition of the Dependable and Dependability
First, taking into account the time in the Commission was devoted to the reach
understanding of the concept of terms dependable and dependability, the prompt
“Define the “dependability” in ICT” was created, looking for definition of the both terms
from AI (“The Chat GPT4.0” and “BingCopilotNotepad” were chosen) perspective. Taking
into account the limitations of the paper, the full prompting history is available at the link
[10].
The ChatGPT4.0 returned rather inclusive definition “In the context of Information and
Communication Technology (ICT), "dependability" is a broad term that encompasses
several key aspects related to the reliability, availability, safety, and security of computer
systems and networks.
Dependability is a measure of a system's ability to deliver a service that can justifiably
be trusted. This concept is crucial in systems where failure or security breaches can lead
to significant consequences, such as in financial transactions, healthcare records, critical
infrastructure, and communication systems.”
In addition, there was provided additional information to the straightforward prompt,
regarding six attributes of “dependability” and explaining each attribute:
“Dependability includes the following attributes:
1.Reliability: The ability of a system to perform its required functions under stated
conditions for a specified period of time without failure. It is often quantified using metrics
such as mean time between failures (MTBF).”
In the similar way next five attributes – availability, safety, security, maintainability
and integrity – were enumerated and explained.
The Bing (CopilotNotepad) returned extensive definition of “dependability” as well: “In
Information and Communication Technology (ICT), "Dependability" is a measure of a
system's availability, reliability, maintainability, and in some cases, other characteristics
such as durability, safety and security.
It refers to the ability of a system to provide services that can be trusted within a time-
period. The service guarantees must hold even when the system is subject to attacks or
natural failures. The International Electrotechnical Commission (IEC) provides systematic
methods and tools for dependability assessment and management of equipment, services,
and systems throughout their life cycles”.
There were six attributes listed as well, but they were not explained, yet there was
something worth making prompt in another AI system: actually working list sources
included, leading to corresponding articles about “dependability” in following sources:
Wikipedia, Springer publishing, Yale University and Study portal.
2.4.2. Comparison of Definitions
When compared with the definition of “dependability” that is found originally in ISO
standard “Dependability 3.1109
1. trustworthiness of a computer system such that reliance can be justifiable placed on
the service it delivers [IEEE 982.1-2005 IEEE Standard Dictionary of Measures of the
Software Aspects of Dependability, 2.2.]
41
2. availability performance and its influencing factors: reliability performance,
maintainability performance and maintenance support performance [ISO /IEC 150256-
1:2013 Systems and software engineering - Systems and software assurance - Part 1:
Concepts and vocabulary, 3.134]”.
When all three definitions are compared (see Table 4), it can be seen that definitions,
provided by ChatGPT4 and Bing clarify the meaning, provided in the ISO standard. In
addition, explanation and sources are given, that helps in understanding of the concept of
the term. The same results were returned for the definition of the term “dependable”.
Table 4
Definitions of “dependability”
ISO ChatGPT BingCopilotNotepad
Dependability 3.1109 Dependability is a In Information and
1. trustworthiness of a measure of a system's Communication Technology
computer system such that availability, reliability, (ICT), "Dependability" is a
reliance can be justifiable maintainability, and in measure of a system's
placed on the service it some cases, other availability, reliability,
delivers [IEEE 982.1-2005 characteristics such as maintainability, and in some
IEEE Standard Dictionary of durability, safety and cases, other characteristics
Measures of the Software security. It refers to the such as durability, safety and
Aspects of Dependability, ability of a system to security. It refers to the
2.2.] provide services that can ability of a system to provide
2. availability performance be trusted within a time- services that can be trusted
and its influencing factors: period. The service within a time-period. The
reliability performance, guarantees must hold even service guarantees must hold
maintainability performance when the system is subject even when the system is
and maintenance support to attacks or natural subject to attacks or natural
performance [ISO /IEC failures. The International failures. The International
150256-1:2013 Systems and Electrotechnical Electrotechnical Commission
software engineering - Commission (IEC) provides (IEC) provides systematic
Systems and software systematic methods and methods and tools for
assurance - Part 1: Concepts tools for dependability dependability assessment
and vocabulary, 3.134] assessment and and management of
management of equipment, equipment, services, and
services, and systems systems throughout their life
throughout their life cycles cycles.
Thus, for the function of better comprehension of the concept of the term, clarifying the
definition both AI systems serves well, at least in the examples examined in this article.
Two next prompts were created with an intention to receive possible terms in source
language that corresponded to the definition: “Using your own definition, please provide
ten most often used terms in ICT, reflecting the meaning, described in the definition” and
“create terms, using given ISO definition”. For further experiment the terms returned to
the ISO definition were chosen (see Table 5).
42
Table 5
Terms returned for ISO definition of “dependability”
ChatGPT4 ChatGPT4 Bing Bing
offered term translation offered term translation
Service Level Uzticamība Reliability Uzticamība
Agreement (SLA)
Reliability: Pieejamība Availability Pieejamība
Availability: Uzturējamība Maintainability Uzturēšana
Maintainability: Uzturējamība Serviceability Servisējamība
Mean Time to Repair Vidējais Mean Time Vidējais laiks
(MTTR) remonta laiks Between Failures starp kļūmēm
(VRL) (MTBF) (MTBF)
Change Management Izmaiņu Fault-Tolerant Kļūdu tolerējošas
pārvaldība Systems sistēmas:
Incident Management Incidentu Data Integrity Datu integritāte
pārvaldība
Disaster Recovery Katastrofu Downtime Neaktīvs laiks
Planning atveseļošanas
plānošana
Failover Mechanism Pārslēgšanās Uptime Darbības laiks
mehānisms
Load Balancing Slodzes Failover Pārslēgšanās
izlīdzināšana
It can be seen that even though there are words that reflect the meaning of the concept
into Latvian, namely “uzticamība”, “darbības laiks”, “pieejamība” and these terms certainly
can serve as an inspiration when looking for a possible terms to be created into Latvian,
there is no ready, sure and reliable answer provided.
The preliminary conclusions for this part of research are as follows: on one hand, AI
systems seem right now useful, when tested in the aspect of providing extensive
definitions, comparing and contrasting different aspects of the terms, and usage examples
in collocations and sentences; on another hand, there is no significant creative aspect in
translation of terms and, what is even more important, there is no guarantee of
repeatability and reliability of the results received when prompting after some time or
different user.
AI works well as a sort of sorting tool for terminology units; yet it is relatively useless
in creating reliable results. Still it can be used as an aide for comparing and contrasting
looking for examples in sentences; looking for synonyms.
Thus, it can be said that the computer program written for sorting the terminology for
meetings is the best short-term solution for speeding up the term creation process.
Sources are indicated, it is reliable, and it reduces the manual labour and provides ground
for fruitful discussion.
43
3. Conclusion
Based on the research work carried out in this article, the following can be
preliminarily concluded.
In the short term, the most efficient way to increase productivity and reduce the time
needed for term pre-processing (preparation for the discussion) is the computer program
created specifically for collecting and exporting already accepted terms in Latvian from
AkadTerm (Academic Term database).
In the long term, there is a possibility that AI systems can be used successfully for
professionals in the STC field for the following specific tasks: a) devising a definition for a
term; b) searching for terms corresponding to the definitions; c) comparing and
contrasting terms and their definitions.
The main drawbacks of AI systems analysed, in order of significance for secondary
term creation, are the unreliability of results, the lack of credibility of sources indicated,
the absence of confidentiality, and the limited amount of term processing.
Further research avenues are a) developing further programs for traceable and
confidential term pre-processing b) specific application of already existing AI systems for
secondary term creation and c) looking into possibilities to train natural language
processing model(s) for term generation purposes.
Acknowledgments
The work described in this paper was supported by the project “Innovative information
technologies” at the University of Latvia. Any errors or omissions are the faults of the
authors.
Disclosure of Interests. The authors have no competing interests to declare that are relevant to the content of this
article.
References
[1] Rusko, T.: English computing terminology as a system, pp. 107–112, (2010), URL:
https://www.researchgate.net/publication/266886183_English_Computing_Termino
logy_as_a_System
[2] Breivik, T.: Action Plan For Norwegian Language And ICT And The Norwegian Human
Language Technology (HLT) Resource Collection, URL: https:// www.sprakradet.no/
localfiles/12399/ifip2005.doc
[3] Cerpja, A., Cepani, A.: Computer Terminology in Albanian Versus other Languages and
the Benefits of Its Codification. Journal of Education and Practice. ISSN 2222-288X.
Vol.5, No.11, 2014, 42-47 (2014). URL: https://www.researchgate.net/
publication/342751546_Computer_Terminology_in_Albanian_Versus_other_Language
s_and_the_Benefits_of_Its_Codification
[4] Diaconu, D.C.: Linguistic Mechanisms Of IT (Computers) Vocabulary Formation In
Romanian Language (2014), URL: https:// www.academia.edu/
44
87984436/Linguistic_Mechanisms_of_It_Computers_Vocabulary_Formation_in_Roma
nian_Language
[5] Darmois, E., Böcker, M.: A Harmonized Multi-lingual Terminology for ICT Devices and
Services with a User-Centric View (2023), URL: https://
www.researchgate.net/publication/ 372227342_ A_Harmonized_ Multi-lingual_
Terminology_ for_ICT_Devices_and_Services _with_a_User-Centric_View
[6] Kavgić, A., Kavgić O.P.: Computer-Related Terminology In Serbian: A Diachronic
Analysis Of Borrowing Trends, Translation And Adaptation Strategies (2011), URL:
https://www.academia.edu/ 2071665/ computer_related_terminology_ in_serbian_
a_diachronic_analysis_of_ borrowing_trends_translation_and_adaptation, last accessed
2024/02/27
[7] Asimov, I.: I, robot. Bantam books, New York (2004)
[8] Clark, A.C.: 2001: A Space Odyssey. ACE, New York (2018)
[9] Sostaka_Borzovs_Zuters_AI_dependable-dependability_DBIS_2024_.pdf, URL: https://
1drv.ms/b/ s!AqrEVf8Pa U0Gg4EgSqoCP6UXp1kL7Q?e=JzqZaH, last accessed
2024/03/20
[10] Sostaka_Borzovs_Zuters_ITTEA_dependable-dependability_DBIS_2024.pdf, URL:
https:// 1drv.ms /b/s!AqrEVf 8PaU0Gg4EhDdMJ74mW1W-XPQ?e=QG4Py5, last
accessed 2024/03/20
[11] Maier, E.: In the Shadows of the ChatGPT Spotlight, URL: http://lt-
innovate.org/content/shadows-chatgpt-spotlight
[12] Nedobity, W.: Terminology and Artificial Intelligence, URL:
https://www.researchgate.net/ publication/327456080_ Terminology_and_
artificial_ intelligence
[13] Rossis, N.C.: The 1960s: A Golden Age for Science Fiction, URL:
https://nicholasrossis.me/2023/08/17/the-1960s-a-golden-age-for-science-fiction/
[14] Science Fiction, URL: https://www.britannica.com/art/science-fiction
[15] Šostaka D., Borzovs J., Zuters J.: How Do We Think? The Case Study of Latvian
Information and Communication Technology Secondary Terms Development, URL:
https://lki.lt/wp-content/uploads/ 2023/09/term_ Vilnius_2023_programme.pdf
[16] AkadTerm, Academic terms database, http://www.akadterm.lv/term.php
[17] Šostaka D., Borzovs J., Zuters J.: Would It Be Possible to Optimise the Secondary ICT
Term Creation in Latvian? 14th Conference on DATA ANALYSIS METHODS for
Software Systems, Vilnius University Proceedings, 89, pp. 1–104.
[18] Rehm, G. et al “The European Language Technology Landscape in 2020: Language-
Centric and Human-Centric AI for Cross-Cultural Communication in Multilingual
Europe” URL: https://www.semanticscholar.org/paper/The-European-Language-
Technology-Landscape-in-2020%3A-Rehm-Marheinecke/02e3c00a7375e4cdd7e654
5b48782aecc0de3420
[19] In 2023, Uba, S.,Y. et all, “A New Computer Science Academic Word List”, 319-330,
URL:
https://www.researchgate.net/publication/369877950_A_New_Computer_Science_A
cademic_Word_List Sciedu Press, World Journal of English Language
45
[20] Izwaini, Sattari in 2005, (ESP Across Cultures 2 (2005) 76-93) in “A corpus-based
study of the translation of IT terms”, URL: https://www.academia.edu/
34913337/Corpus_based_Study_of_IT_Terms.
[21] Postolea, Sorina– 2011 - Hybrid Terms in the Romanian Language of Computer
Science URL: https://www.academia.edu/5200765/Hybrid_Terms_in_the_Romanian_
Language_of_Computer_Science
[22] Postolea, Sorina– 2012, from regularities to norms in the secondary formation of
romanian ict terms, linguaculture, 2, 2012, 101-115., secondary formation of
Romanian ICT terms URL: https://www.researchgate.net/publication/276172400_
From_Regularities_to_Norms_in_the_Secondary_Formation_of_Romanian_ICT_Terms
[23] Postolea, Sorina– 2014, The Phraseology of General Use ICT Texts and its Translation
from English into Romanian Volumul X, Nr. 2, iunie 2014 URL:
https://www.researchgate.net/publication/ 281589550_ The_Phraseology_of_
General_Use_ICT_Texts_and_its_Translation_from_English_into_Romanian_-_A_ Corpus
-based_Analysis
[24] Artificial Intelligence for the Humanities. Fundamental Concepts Explained (2024)
URL: https://www.researchgate.net/publication/378335852_Artificial_Intelligence_
for_the_Humanities_Fundamental_Concepts_Explained
[25] Corpus linguistics and generative AI tools in term extraction: a case of Kashubian - a
low-resource language (2023) URL: https://www.researchgate.net/publication/
378234423_Corpus_linguistics_and_generative_AI_tools_ in_term_ extraction_ a_case_
of_ Kashubian_-_a_low-resource_language
[26] Terminology work in the era of AI-generated content. A case of ChatGPT (an end-user
perspective) URL: https://www.researchgate.net/publication/ 378297882_
Terminology_work_in_the_era_of_AI-generated_content_ A_case_of_ ChatGPT_an_end-
user_perspective
[27] A Trilingual Creation of Terms of the Computer Language in English, French and Igbo,
LALIGENS, VOL. 7(2), S/N 16, OCTOBER, 2018, 37-48. URL:
https://www.researchgate.net/ publication/ 329098902_ A_trilingual_ creation_ of_
terms_ of_the_computer_language_in_English_French_and_Igbo
[28] Meiramova Saltanat Akimovnaa*, Mussagozhina Kuanysh Kanatovnab, The
investigation of English, Russian and Kazakh computer terms borrowings to be
acquired at English class, 94-102, URL: https://www.sciencedirect.com/
science/article/pii/S187704281504495X
[29] Šostaka D., Borzovs, J, Cauna. E., Keiša I., Pinnis M., Vasiļjevs A.: The Semi–Algorithmic
Approach to Formation of Latvian Information and Communication Technology
Terms Baltic J. Modern Computing, Vol. 11 (2023), No. 1, 67-89 URL:
https://doi.org/10.22364/bjmc.2023.11.1.05
[30] Šostaka D., Borzovs, J.: Towards Computer-Assisted Latvian ICT Terminology
Development: from Theoretical Guidelines to Case Studies, Baltic J. Modern
Computing, Vol. 11 (2023), No. 1, 134-154, URL:
https://doi.org/10.22364/bjmc.2023.11.1.08
46
[31] Šostaka D., Borzovs J., Zuters J.: Preliminary Results of the Use of Artificial Intelligence
in Secondary Term Formation. Digital Humanities, 2024-04-25, URL:
https://drive.google.com/file/d/12a3lZpvv7SxzBo2B03zu1fhdePinCy9n/view
[32] Kageura K., and Umino B.: Methods of automatic term recognition: Terminology.
International Journal of Theoretical and Applied Issues in Specialized Communication,
Volume 3, Issue 2, Jan 1996, p. 259 – 289 DOI:
https://doi.org/10.1075/term.3.2.03kag
[33] Fulford H.: Exploring terms and their linguistic environment in text: A domain-
independent approach to automated term extraction. Terminology. International
Journal of Theoretical and Applied Issues in Specialized Communication, Volume 7,
Issue 2, Jan 2001, p. 259 – 279, DOI: https://doi.org/10.1075/term.7.2.08ful
[34] Serrec A. L., L'Homme Marie-Claude, Drouin P. and Kraif O.: Automating the
compilation of specialized dictionaries: Use and analysis of term extraction and lexical
alignment. Terminology. International Journal of Theoretical and Applied Issues in
Specialized Communication, Volume 16, Issue 1, Jan 2010, p. 77 – 106 DOI:
https://doi.org/10.1075/term.16.1.04les
[35] Gurbanova A. M.: Analysis of automation methods of the term creation activity. 2018.
Problems of Information Technology. URL: https://jpit.az/en/journals/219
[36] Schulz S., Hammer L., Hashemian-Nik D., Kreuzthaler M.: Localising the Clinical
Terminology SNOMED CT by Semi-automated Creation of a German Interface
Vocabulary. Proceedings of MultilingualBIO 2020, pages 15–20. URL:
https://aclanthology.org/2020.multilingualbio-1.3.pdf
47