Developing an Artificial Intelligence Model to Enhance the Emotional

Intelligence of Motor Vehicle Drivers for Safer Roads

Ana Todorova

, Irina Kostadinova

and Svetlana Stefanova

University of Ruse “Angel Kanchev”, 8 Studentska Street, Ruse, Bulgaria

Keywords: Emotional Intelligence, Artificial Intelligence, Virtual Reality, Training, Safety Roads.

Abstract: Traffic accidents and risky driving behaviour are among the deadliest problems worldwide. This statement

becomes an undeniable fact, thanks to the grim statistics of the World Health Organization, according to

which more than 1 million people die on the roads every year. Road accidents are also among the leading

causes of death among children and young people aged 5 to 29. Against this background, a number of studies

look for a link between the emotional intelligence of motor vehicle drivers and the potential prevention of

risky driving. Building on the scientific knowledge generated up to this point, the present study suggests a

prototype of an AI-based model that aims, through ongoing assessment and subsequent training, to enhance

the emotional intelligence of both future and current motor vehicle drivers who are prone to risky behaviour

on the road. Through simulated scenarios in a virtual environment, the model aims to improve the ability of

drivers to recognise and manage their own and other people's emotions and to react adequately to different

situations on the road. The expectation is that the model will reduce the manifestations of aggression and

intolerance on the road and ultimately lead to safer roads.

1 INTRODUCTION

Risky driving is a recognised factor in road traffic

accidents. Driving behaviour significantly influences

the occurrence of traffic accidents and fatalities.

Intentional dangerous behaviours, such as exceeding

speed limits or driving under the influence of

substances, are among the predominant contributing

reasons for traffic accidents. It is estimated that

between 90% and 95% of traffic accidents worldwide

are the result of human error (Aniah, 2021; Ahmed et

al., 2022). Given the serious and often fatal outcomes

associated with risky driving practices, certain

behaviours can seem confusing when viewed from a

rational perspective. In this context, theories of risky

decision-making emphasise the influence of emotions

on an individual's actions in dangerous situations

(Megías-Robles et al., 2022).

Emotionally agitated drivers may approach

hazards with impaired attention and thus –

unintentionally or not – engage in reckless driving.

Therefore, the tools by which drivers can exercise

https://orcid.org/0009-0007-2993-077X

https://orcid.org/0000-0001-8845-7598

https://orcid.org/0000-0003-4500-4833

control over their own emotions and thus prevent

risky driving are vital to getting drivers safely from

point A to point B. Emotional control over oneself, as

well as over others, is imperative in driving situations

(Ahmed et al., 2022).

According to Megías-Robles et al. (2022), the

driver's emotional state is a critical factor in

explaining risk-taking propensity. The authors also

claim that an adequate ability to perceive, understand,

and manage emotions would allow drivers to have

better control over their emotional condition and their

perception of other road traffic participants. As a

result, it would help to reduce participation in risky

behaviours and the number of road accidents.

The relationship between emotional intelligence

(EI) and driving behaviour is the subject of increasing

interest in the aspect of risky and aggressive driving

behaviour, especially in the context of young drivers

(Aniah, 2021). Drivers' emotions have been found to

be among the main factors contributing to dangerous

driving behaviour. Emotions can be measured,

understood and regulated most effectively through EI.

Todorova, A., Kostadinova, I. and Stefanova, S.

Developing an Artiﬁcial Intelligence Model to Enhance the Emotional Intelligence of Motor Vehicle Drivers for Safer Roads.

DOI: 10.5220/0013051400003838

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 16th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2024) - Volume 3: KMIS, pages 339-346

ISBN: 978-989-758-716-0; ISSN: 2184-3228

339

To varying degrees, large-scale studies have

confirmed the link between EI and dangerous driving

behaviour (Ahmed et al., 2022). Emotion affects

drivers due to its influence on the degree of self-

control, thus also affecting the driving method

(Aniah, 2021). A 2019 study also found that people

who drive every day have poor EI, which hinders safe

driving. The study suggests that training drivers in

emotional regulation can contribute to safer roads

(Parameswaran and Balasubramanian, 2020).

Taking into account the studies already carried out

on the subject, the present study aims to develop an

algorithm that will help develop the EI of vehicle

drivers. The report is structured as follows: 1)

Literature review of studies looking for a link

between EI and driving; 2) Presentation of the

research methodology; 3) Presentation of the

developed model for evaluating and increasing EI; 4)

Conclusion.

2 THEORETICAL

BACKGROUND

At the end of 2023, the World Health Organization

announced that around 1.19 million people die

worldwide due to road traffic accidents annually.

Road traffic accidents alone are the leading cause of

death for children and young people between 5 and

29 years of age (WHO, 2023).

Similarly, in the United States, the National

Highway Traffic Safety Administration states that in

2018, more than 2.7 million people were injured, and

36,096 people died in motor vehicle crashes. The data

are frightening and show beyond doubt that it is

imperative to identify the factors contributing to road

accidents and to take measures to reduce them. Such

factors may include vehicle defects and

environmental obstructions, i.e. road and weather

conditions, but dangerous driving behaviour also

emerges as a significant factor. The latter is defined

as any inappropriate driver activity that increases road

hazards and the likelihood of a vehicle crash (Ahmed

et al., 2022).

At the same time, drivers engage in dangerous

driving for a variety of reasons, including fatigue,

distraction, and driving under the influence of alcohol

or drugs. A number of psychological factors, such as

personality traits and emotions, have also been found

to contribute to dangerous driving behaviour (Owsley

et al., 2003). Research has alerted to the fact that

drivers' emotions can particularly strongly influence

their destructive driving behaviour. Previous research

has also emphasised the correlation between different

emotions (e.g., feelings of frustration, anger, and

sadness) and aggressive driving (Ahmed et al., 2022).

According to Aniah (2021), physiological and

psychological variables such as gender, experience,

age, and emotions inevitably influence the behaviour

of drivers. The author defines driving as a

psychomotor ability because it consists of body

movement and a cognitive task. This ability can take

many forms, but the technique a driver uses depends

on their personality and behavioural profile.

In a broad sense, EI is a construct that

encompasses all of a person's emotional abilities. In

this regard, a number of studies have sought and

found a relationship between EI and risky driving

behaviour (Ahmed et al., 2022). The results are of

varying degrees of certainty due to the existence of

different approaches to EI in the literature. The self-

report ability model understands EI as a mental ability

and focuses on the emotional skills included in a

conceptualisation of EI. It uses self-report measures

to assess these abilities. The performance-based

ability model also views EI as a set of emotion-related

abilities but assesses them using performance-based

tests. Finally, the mixed model conceptualises EI as a

broad construct combining both emotion-related

skills and personality factors that are evaluated

through self-report instruments (Megías-Robles et

al., 2022).

The primary motivation to look for a link between

EI and driving is that it is the driving style of many

drivers that is responsible for the significant number

of accidents that occur (Aniah, 2021). The results of

a study by Megías-Robles et al. (2022) found that

higher self-reported EI, especially the ability to

regulate emotions, was associated with a lower

propensity for risky driving. According to the authors,

emotion regulation and evaluation of the emotions of

others are EI abilities that may predict the number of

potential accidents. EI, which is explained as a

person's ability to recognise, identify, use, express, as

well as regulate their own and others' emotions, has

been empirically proven to influence driving

behaviour (Aniah, 2021).

Such a conclusion is hardly surprising to anyone.

Emotions are a fundamental part of human behaviour

– they guide an individual's attention, memory,

motivation, and even decision-making process. In

risky contexts, however, emotions are essential, given

the time (momentary) pressure and substantial

emotional consequences these situations often

involve. The integration of emotional factors in the

processing of risky behaviour has also been

demonstrated at the neural level, including in the

KMIS 2024 - 16th International Conference on Knowledge Management and Information Systems

340

context of driving (Hernandez et al., 2014). Driving

is an activity in which emotions often arise – traffic

jams, accidents, risky traffic participants, etc. – all

these situations can cause fear, as well as retaliatory

aggression, intolerance, and dangerous behaviour. In

many cases, these emotions underlie human

behaviour, but their consequences in risky driving are

particularly significant (Megías-Robles et al., 2022).

There is no doubt that emotion can arise at any

particular moment while driving and can have a

different emotional impact on the driver's behaviour.

Therefore, various emotional states can affect driving

differently because people differ in how they react to

situations. This makes emotional control of self and

others imperative in driving conditions (Owsley et al.,

2003).

Aniah (2021) opined that driving behaviour is

usually a pattern chosen by the driver himself.

Therefore, it is argued that the specific style and skills

that a motor vehicle driver applies at a given place

and time are strongly influenced by his emotions and

the relationship between stimulus and response. The

link between stimulus and response, or rather the

bridge between them, represents the individual's EI.

The above gives reason to generalise that drivers

with reported higher EI scores engage in less

dangerous driving, which is reflected in fewer crashes

and fatalities. Therefore, as reported by Ahmed et al.

(2022), promoting and improving EI can be helpful in

preventing risky driving among non-professional

drivers. Incorporating EI training into driver training,

on-the-job training, and licensing procedures can help

develop safer drivers.

The problem with this type of evaluation of EI and

training is the human factor, i.e. the private interest or

the subjective opinion of the trainer, which can be a

prerequisite for intentional or unintentional mistakes

and corrupt practices in such an essential field as road

safety. With the help of large language models and

the incorporation of AI during the learning stage of

driving for young adults, it can be very beneficial for

them to be assessed and trained based on their EI.

Their behaviour can be evaluated and then compared

to “ideal” behaviour, which can help them nip the

negative traits in the beginning stages before passing

their driving test and becoming experienced drivers.

Such negative behaviour is challenging to be

subjected to change, if not impossible, after years and

years of driving experience.

That motivated the authors of the present study to

develop a prototype model based on AI that would

balance, without removing, the human factor and

minimise the associated risks. In addition, a central

place in the developing AI model is precisely the

emotions and, more specifically, the ability of the

individual to recognise (empathy) and regulate

(self/control) both his own and others' emotions. The

key thing about these two elements is that, although

worded differently, they are involved in the different

models of EI, which makes them generalisable to the

EI personality.

3 METHODOLOGY

The methodology of the study includes three main

stages: 1) Literature review (introduced in the section

"Theoretical background"), 2) Determination of

assessing variables (described in the section

"Methodology", and 3) Explanation of the

construction of the developing AI model (described

and analysed in "Results" and "Discussion" sections):

elements selection, model structure, action algorithm,

process analysis.

Based on the theoretical analysis, the research

methodology introduces three basic variables, which,

with the help of information and communication

technologies (ICT) and human control, will be

prioritised for research, analysis and improvement:

emotion regulation or self-control (x), emotion

recognition (in others and self) or empathy (y), and

evaluation of the EI as a construct (z).

The measurement and evaluation of variable z

will follow one of the three schools of EI: cognitive

ability, personality trait, or mixed model. The choice

is yet to be refined and validated. Based on the

selected overall construct, additional variables will

be introduced, such as self-knowledge or social skills

(elements of the mixed EI model).

The need to introduce baseline variables in the

first place is justified by existing research on the

subject of the relationship between the baseline

variables specified and the driving pattern applied by

the individual. Implementation of additional variables

will look for underestimated or so far neglected

correlations between EI components and driving

behaviour. At this stage, these variables are not built

into the model and its algorithm.

The developed AI model is tied to risky driving,

which is why the algorithm will look for correlations

between the basic and potential additional variables

introduced so far and the characteristics of the motor

vehicle driver identified in the background literature,

influencing the genetic driving style. These key

variables are: aggression (a), intolerance (b), and

risk-taking (c).

An interdisciplinary team of Bulgarian university

scientists with expertise in management, risk

Developing an Artiﬁcial Intelligence Model to Enhance the Emotional Intelligence of Motor Vehicle Drivers for Safer Roads

341

management, entrepreneurship, social sciences, and

ICT participated in developing the methodology and

AI model. The goal is to specify as much as possible

the main elements of the model, the algorithm, and

the overall process from development to

implementation and validation of the AI model.

4 RESULTS

From the perspective of the ICT that build the model's

algorithm, the research introduces a toolkit of three

technologies: artificial intelligence (AI), virtual

reality (VR), and blockchain technology (BT). Table

1 describes the need for their applicability.

Table 1: Type and applicability of implemented ICT.

Source: own development.

Technology

Applicability

Artificial

Intelligence (AI)

Definition: It focuses on the creation

of intelligent agents, i.e., systems

capable of perceiving their

environment, reasoning, and acting

to achieve specific goals. AI models

strive to mimic and sometimes even

surpass human cognitive abilities

such as learning, problem-solving,

pattern recognition, and natural

language understanding (Antonova

et al., 2021; Gignac & Szodorai,

2024).

Applicability: AI models can

analyse large amounts of data faster

and more accurately than human

raters. At the same time, in big data

sets, AI models can detect complex

patterns that are difficult or even

impossible for humans to spot. In

addition, while there is still debate

about their accuracy, AI models also

offer greater objectivity. They are

less susceptible to subjective bias,

which reflects in more objective

results. These two advantages, along

with automating the evaluation

process and saving time and

resources, make AI models an

adequate substitute for human

resources.

Virtual reality (VR)

Definition: VR is a technology that

creates immersive and interactive

environments that simulate real or

imagined scenarios. VR users wear

headsets that display 3D images and

sounds and sometimes use

controllers or gloves to interact with

the virtual world. VR can create

realistic and engaging experiences

that can stimulate emotions,

thoughts, and behaviours (Susindar

et al., 2019).

Applicability: Well-developed EI is

largely believed to result from an

individual's experience, which is

why parallels between EI and

wisdom are often sought and found.

Through VR technology, the

participant will be placed in

situations that mimic real life and

are known to help develop EI

competencies.

Blockchain

Technology (BT)

Definition: A distributed database

that records transactions in a way

that is secure, transparent, and

resistant to change. Information in

the blockchain is organised into

blocks that are linked together by

cryptographic hashes. Once added,

information on a blockchain is

virtually impossible to change,

making it highly reliable. This

technology is decentralised, meaning

there is no one central authority to

control the network. Instead,

blockchain networks are managed

by multiple computers, making them

resistant to censorship and

manipulation (Tripathi et al.,

2023).

Applicability: Implementing the BT

on the AI model is necessary to

preserve the confidentiality of each

training participant's data.

Personality test results do not

constitute medical information by

themselves. However, the fact that

they provide personal and sensitive

information for the individual

requires a more serious commitment

to their preservation.

The developed model for increasing the EI of

drivers is based on three AI models: machine

learning, deep learning and natural language

processing, and more specifically:

1) Machine Learning: Classification, Regression,

Clustering

The classification will serve to classify participants'

responses into different personality types or traits,

such as "extroverted" or "introverted." Regression

will be applied to predict the values of the entered

(baseline, additional, and key) variables related to the

learner's personality. Clustering will be applicable

when grouping participants with similar personality

profiles.

KMIS 2024 - 16th International Conference on Knowledge Management and Information Systems

342

2) Deep Learning: Neural Networks, Recurrent

Neural Networks

Neural networks will simulate the workings of the

human brain and can detect complex patterns in data.

They are used for natural language analysis, pattern

recognition and processing large amounts of

unstructured data. Recurrent neural networks, on the

other hand, are suitable for analysing sequential data,

such as text responses to questions in personality

tests. They can pick up on contextual information and

extract deeper meanings.

3) Natural Language Analysis: Sentiment

Analysis, Keyword Extraction, Semantic

Content Analysis, Personalised

Recommendations

AI models will be used to analyse large amounts of

text responses to personality test questions to identify

keywords, phrases and emotional responses that are

characteristic of certain personality traits. In some

cases, AI models will operate to analyse facial

expressions, gestures and other non-verbal signals to

gain additional information about an individual's

personality. As a result, AI will create personalised

recommendations for an individual's EI development

based on test results, as well as for the virtual

experiences that will be most relevant to the object,

subject and purpose of the learning.

Figure 1 presents the algorithm of the developed

AI model. The whole process is divided into four

main Stages: I) Measurement and evaluation of the

introduced basic (x, y, z), key (a, b, c) and additional

variables; II) Verification of the results obtained by a

person; III) and IV) Depending on the results of Stage

II – respectively unsatisfactory or satisfactory from

the point of view of human evaluation, in the next

stage either a human specialist makes a new

measurement and evaluation of the variables, or the

AI determines the virtual experience that the object to

be subjected to.

The model introduces four working agents: (A) –

learner, (B) – training/evaluating AI, (C) –

training/assessing/supervising specialist persons, and

(D) – AI-defined the virtual experience. In Stages I

and III, blockchain encryption of the data generated

from the tests to measure the sought variables was

introduced. This, as stated at an earlier stage, is

imperative from the point of view of the object's

confidentiality and to protect its data. The results of

such tests, if publicly available, may have an adverse

impact on an individual's personal and professional

development in the future. Also, blockchain

technology will neutralise the possibility that the

records of the results can be tampered with and

manipulated.

The role of virtual reality, on the other hand, is to

act as an imaginary learning environment, where

learning does not mean learning phrases or

behavioural responses (Han & Lorenzo Najord,

2024). On the contrary, the goal is a rapid

accumulation of life experience, but in a protected

(virtual) environment, with this experience directly

related to developing the basic variables – self-

regulation and empathy, and reducing the key

variables – aggression, intolerance, propensity to risk.

In fact, in virtual reality, the individual will be trained

to respond most effectively to the stimulus-response

interdependence that Aniah (2021) also talks about.

This effective response is due to a developed EI

capability.

Figure 1: Working algorithm of the developed AI model for

evaluating and improving the EI of motor vehicle drivers

for safer roads. Source: own development.

In their study, Susindar et al. (2019) demonstrate

that the use of VR can be an effective emotion-

inducing method when investigating emotion-

Developing an Artiﬁcial Intelligence Model to Enhance the Emotional Intelligence of Motor Vehicle Drivers for Safer Roads

343

influenced decision-making. As stated in the

theoretical background, this is highly inherent to

drivers of motor vehicles who are daily exposed to

emotionally arousing situations and have to make

immediate decisions and respond to the stimulus-

response relationship according to their momentary

mood. Susindar and his co-authors' research focused

on extracting and generating situations that evoke fear

and anger – emotions that drivers also face on the

road. At the same time, the authors of the cited study

add that it is not entirely clear how the virtual

environment affects performance (or learning) and

the degree to which emotions are evoked. This means,

on the one hand, that it is not categorically clear

whether an individual would react in the same way in

a real and virtual environment and, on the other hand,

that more in-depth research is needed in this direction.

However, the latter in no way belittles VR technology

capabilities, as evidenced by similar studies on the

topic (Marques et al., 2022; Mancuso et al., 2023;

Hariyady et al., 2024).

Considering the rapid growth of technology in

every aspect, people are more and more concerned

about how to improve user experience rather than the

construction of the experience itself. Giving people

the opportunity to experience this in virtual reality

would allow them to see their mistakes and correct

them in their free time in a safe environment, making

them better on the road without the risks of actually

being there. This would help them improve their EI

time and time again, improving the quality of their

driving skills, thus improving the overall quality of

driving for the other drivers on the road around them.

As the graph (Figure 1) shows, the process does

not stop after the learner's experience in virtual

reality. New measurements and evaluation of the

basic variables, as well as the additional and key

variables adopted in the Methodology, are needed.

Therefore, the algorithm starts again from Stage I.

The subsequent steps, i.e. whether to stop or continue

training, depend on the evaluation of agents B and C

and the decision of agent B. Realistically, the process

of increasing EI may never stop and even apply to

refresher driving courses (similar to the periodically

conducted psychological tests) of experienced drivers

of motor vehicles.

Although still unproven categorically, some EI

researchers believe that this ability and its inherent

competencies have the potential to develop positively

over time, one of the reasons being precisely the

accumulated life experience. The AI model

developed and presented in the present study is based

precisely on the statement that EI can be developed

and that it undergoes evolution with age (Gilar-Corbi

et al., 2019).

5 DISCUSSION

When analysing the algorithm, the logical question

arises: What necessitates human monitoring and

control of the evaluation of the AI model?

Conversely, if this activity is entirely within a

human's capabilities, why is it necessary to introduce

the use of AI?

First of all, the potential of AI to process vast

amounts of data at a speed beyond the reach of a

human being has already been repeatedly tested and

proven. However, not only is speed essential, but so

is the refinement of the results. Subjectivity and

unintentional omissions in data analysis are human.

But the same goes for imperfect (yet) AI, which, like

its creator, is also not immune to biases and errors in

its algorithm. In fact, the AI model being developed

to increase the EI of drivers shows precisely how

natural and artificial intelligence should collaborate –

in a balanced synergy, but also with a good-natured

mistrust of the abilities of one or the other.

On the other hand, it is essential to note that AI

models cannot and are not expected to replace human

raters. In the overall model, the human factor is equal

to the involvement of AI. This further justifies the

intervention of blockchain technology in the

algorithm. The encryption of the data from the

conducted tests and training will guarantee the

confidentiality and objectivity of the process.

Therefore, AI models should be used as an additional

tool to justify and support decision-making. To

ensure the validity and reliability of AI-based test

results, careful validation studies need to be

conducted.

Attention to detail throughout the process is

significant as irreparable damage can be done to the

learner's psyche, which is not the purpose of the AI

model. It is extremely crucial to create and apply the

"training" in virtual reality according to the test

results of the entered variables. A hypersensitive

person would accept such experiences in one way,

and a more selective individual in his feelings would

have a radically different perception of what is

happening. Therefore, every AI-proposed and

human-approved virtual learning environment in EI

must be perfect and maximally adapted to the

personal qualities of the respective learner.

It is equally important to explain why the basic

variables, such as empathy and self-regulation, are set

in the algorithm rather than examining only the EI

KMIS 2024 - 16th International Conference on Knowledge Management and Information Systems

344

construct as a whole. Or why the general construct EI

is asked at all, and not only the connections between

empathy and self-regulation of motor vehicle drivers

and their aggressive and angry behaviour that puts

them in risky situations are investigated. As stated

earlier, self-regulation and empathy are part of the

general EI construct. Although in the pilot presented

AI model, as well as numerous other studies, they are

studied as independent competencies, in reality, it is

not possible to fully develop these two abilities

without the other EI competencies, regardless of

which of the three EI models will be followed. For

example, the mixed model includes self-regulation

and empathy, as well as self-knowledge, motivation,

and social relationships. Goleman himself, one of the

creators of the model, points out that it is not possible

to achieve self-regulation without self-awareness

since each of the sub-competencies of self-

management steps on self-awareness (Goleman and

Chernish, 2023). This necessitates the study,

evaluation and improvement of the general EI

construct by examining the construct's constituent

components separately, i.e., it is crucial to approach

the problem deductively.

Limitations of the Study

A significant limitation of the study and the

introduced model is the lack of definitive data on the

effectiveness of some of the applied ICT. The

analyses and tests carried out so far are scarce, and

the results are contradictory. Therefore, the

application of the described technologies does not

guarantee the desired result.

Another limitation of the study is the lack of data

from the practical application of the developed

model. The presented model is purely theoretical and

has not been validated empirically, which limits the

ability to conclude its reliability and predictive

validity. To overcome this limitation, it is necessary

to conduct an empirical study in which the model is

tested on a large and heterogeneous sample of

participants. The formed interdisciplinary team needs

to conduct considerable research and testing in this

direction, but the model aims to make a start.

On the other hand, implementing such an

algorithm in a seemingly state-controlled but

apparently private-interest-dominated environment,

such as driving schools, requires considerable will

and agreement from multiple (dis)interested parties.

In this sense, just creating the AI model is not enough.

Therefore, the author collective's future efforts will be

directed to the experimental introduction of the model

and the search for validation of the algorithm and the

expected results.

Applicability

The goal of the research team is to apply the algorithm

as a priority in the courses for acquiring driving skills.

It is an environment in which the individual most

actively reveals himself – as more aggressive, more

intolerant, more selfish or vice versa. Bulgaria is first

in the EU in terms of road deaths in 2023. The

European Commission published preliminary data for

last year – the EU average is 46 deaths per million

inhabitants. In Bulgaria, the ratio is 82 victims per one

million inhabitants (Apostolova, 2024). In this sense,

the developed AI model is expected to lead to

significantly greater self-awareness and self-

regulation of their own emotions in both

inexperienced and seasoned drivers on the roads.

At the same time, the developed AI model can

also be used in a number of other areas where weak

self-regulation of emotions leads to conflict situations

or where reducing the levels of aggression is

necessary. The results of the implementation will

validate the results and allow the model to be

embedded in educational programs and its application

in the fight against hate speech, intolerance of

differences, selfishness, cruelty to the weaker, etc.

This will be the subject of future research by the

authors.

6 CONCLUSION

Death on the roads as a result of serious road traffic

accidents is one of the saddest facts of our time, which

we must either accept or overcome. Over 1 million

people die on the road, not a small number of them

children and young people. The purpose of the

developed AI model to increase the EI of drivers is

precisely this – fighting statistics, risky driving, but

above all, saving lives before they are even in danger.

It is essential to clarify that in this "battle", neither

EI nor AI is a panacea. The human factor remains a

significant unknown, along with the cultural and

social characteristics of one or another country.

However, suppose it is almost impossible to change

the cultural and social conditions. In that case, only

the way in which the individual responds to the

stimulus-response relationship is within his

capabilities.

In conclusion, the developed AI model is a small

step towards achieving high EI drivers of motor

vehicles and reducing risky behaviour on the road.

Although further research and efforts are needed to

implement this technology, it is essential to find a

way to save human lives and reduce the socio-

economic costs directly related to traffic accidents.

Developing an Artiﬁcial Intelligence Model to Enhance the Emotional Intelligence of Motor Vehicle Drivers for Safer Roads

345

ACKNOWLEDGMENTS

This study is financed by the European Union-

NextGenerationEU through the National Recovery

and Resilience Plan of the Republic of Bulgaria,

project №BG-RRP-2.013-0001-C01.

REFERENCES

Ahmed, J., Ward, N., Otto, J., McMahill, A. (2022). How

does emotional intelligence predict driving behaviors

among non-commercial drivers? Transportation

Research Part F: Traffic Psychology and Behaviour,

Volume 85, 38-46, February 2022. DOI: 10.1016/j.trf.

2021.12.013.

Aniah, E. (2021). The influence of Emotional Intelligence

on Driving Behavior in the Transportation Sector: A

literature review. Journal of Contemporary Issues in

Business and Government, Volume 27, Issue 06, 1524-

1537, December 2021. DOI:10.47750/cibg. 2021.27.

06.121.

Antonova, D., Kunev, S., Venelinova, N., Kostadinova, I.

(2021). Data Management Conceptual Algorithm of

Transnational Digital Scientific Infrastructure as an

Answer to the Globalization Challenges, 20th

International Scientific Conference Globalization and

Its Socio-Economic Consequences University of Zilina,

1-9, DOI: 10.1051/shsconf/20219205002.

Apostolova, D. (2024). България на първо място в ЕС по

катастрофи, завършили със смърт [Bulgaria ranks

first in the EU in fatal accidents], BNT 1, 8 March 2024.

https://bntnews.bg/news/balgariya-na-parvo-myasto-v-

es-po-katastrofi-zavarshili-sas-smart-1269911news.

html [In Bulgarian] (Accessed 16 August 2024).

Gignac, G., Szodorai, E. (2024). Defining intelligence:

Bridging the gap between human and artificial

perspectives. Intelligence, Volume 104, 101832, May–

June 2024. DOI: 10.1016/j.intell.2024.101832.

Gilar-Corbi, R., Pozo-Rico, T., Sánchez, B., Castejón, J.

(2019). Can emotional intelligence be improved? A

randomized experimental study of a business-oriented

EI training program for senior managers. PLoS One,

Volume 14, Issue 10, e0224254, 23 October 2019. DOI:

10.1371/journal.pone. 0224254.

Goleman, D., Chernish, C. (2023). Optimal: How to Sustain

Personal and Organizational Excellence Every Day,

Publisher: Harper Business.

Han, M., Lorenzo Najord, I. (2024). Virtual reality with e-

learning model in business management and knowledge

complex environment for gaming analysis. Entertain-

ment Computing, Volume 50, 100703, May 2024. DOI:

10.1016/j.entcom.2024.100703.

Hariyady et al. (2024). Virtual Reality and Emotional

Responses: A Comprehensive Literature Review on

Theories, Frameworks, and Research Gaps. ITM Web

of Conferences 63, 2-13, 01022, February 2024. DOI:

10.1051/itmconf/20246301022.

Hernandez, J., McDuff, D., Benavides, X., Fernandez, J.,

Maes, P., Picard, R. (2014). AutoEmotive: Bringing

empathy to the driving experience to manage stress.

DIS Companion '14, 21-25 June 2014, Vancouver, BC,

978-1-4503-2903-3/14/06.

DOI:10.1145/2598784.2602780.

Mancuso, V., Bruni, F., Stramba-Badiale, C., Riva, G.,

Cipresso, P., Pedroli, E. (2023). How do emotions

elicited in virtual reality affect our memory? A

systematic review. Computers in Human Behavior,

Volume 146, 107812, September 2023. DOI: 10.1016/

j.chb.2023.107812.

Marques, A. J., Gomes Veloso, P., Araújo, M., de Almeida,

R. S., Correia, A., Pereira, J., Queiros, C., Pimenta, R.,

Pereira, A. S., Silva, C. F. (2022). Impact of a Virtual

Reality-Based Simulation on Empathy and Attitudes

Toward Schizophrenia. Frontiers in psychology,

Volume 13, 814984, 4 May 2022. DOI: 10.3389/

fpsyg.2022.814984.

Megías-Robles, A., Sánchez-López, M., Fernández-

Berrocal, P. (2022). The relationship between self-

reported ability emotional intelligence and risky driving

behaviour: Consequences for accident and traffic ticket

rate. Accident Analysis & Prevention, Volume 174,

September 2022. DOI: 10.1016/j.aap.2022.106760.

Owsley, C., McGwin Jr., G., McNeal, S. (2003). Impact of

impulsiveness, venturesomeness, and empathy on

driving by older adults. Journal of Safety Research,

Volume 34, Issue 4, 2003, 353-359. DOI: 10.1016/

j.jsr.2003.09.013.

Parameswaran, S., Balasubramanian, V. (2020). Assessing

the Relation Between Emotional Intelligence and Dri-

ving Behavior: An Online Survey. Advances in Human

Factors of Transportation, Volume 964, 233–239, 06

June 2019. DOI: 10.1007/978-3-030-20503-4_21.

Susindar, S., Sadeghi, M., Huntington, L., Singer, A.,

Ferris, T. (2019). The Feeling is Real: Emotion

Elicitation in Virtual Reality, Proceedings of the

Human Factors and Ergonomics Society Annual

Meeting, Volume 63, Issue 1, 252-256, 20 November

2019. DOI:10.1177/1071181319631509.

Tripathi, G., Ahad, M. A., Casalino, G. (2023). A

comprehensive review of blockchain technology:

Underlying principles and historical background with

future challenges. Decision Analytics Journal, Volume

9, 100344, December 2023. DOI: 10.1016/j.dajour.

2023.100344.

WHO (2023). Global status report on road safety 2023.

World Health Organization, 13 December 2023, ISBN:

978-92-4-008651-7. https://www.who.int/publications/

i/item/9789240086517 (Accessed 16 August 2024).

KMIS 2024 - 16th International Conference on Knowledge Management and Information Systems

346