Article citation information:
Czech, P.
Artificial intelligence as a basic problem when implementing autonomous vehicle
technology in everyday life. Scientific Journal of Silesian University
of Technology. Series Transport. 2024, 122, 49-60. ISSN: 0209-3324. DOI: https://doi.org/10.20858/sjsutst.2024.122.3.
ARTIFICIAL INTELLIGENCE AS A BASIC PROBLEM WHEN IMPLEMENTING AUTONOMOUS
VEHICLE TECHNOLOGY IN EVERYDAY LIFE
Summary. Innovative
technologies that use artificial intelligence in transport solutions recently
emerging around the world include, among others issues of autonomous vehicle
driving. The use of autonomous vehicle technology affects the issues of civil
liability (liability and insurance), road safety, natural environment (energy
efficiency, renewable energy sources), data (access, exchange, protection,
privacy), IT infrastructure (effective and reliable communication), employment
(creation and loss of jobs, training of truck drivers in the use of automated
vehicles). The development of new technologies related to artificial
intelligence, including autonomous vehicles, generates inevitable changes in
law, economy and society. It is inevitable due to the fact that autonomy is
undoubtedly a means to achieve the goal of improving the efficiency sought in
every area of life. The article presents arguments confirming the thesis that
the basic factor inhibiting the implementation of autonomous vehicle technology
is the problem of artificial intelligence, including its definition and legal
regulation.
Keywords: artificial
intelligence, autonomous vehicles, legal regulations, innovative technologies,
mobility, transport
1.
INTRODUCTION
In
his actions, man has always tried to set new challenges for himself, forcing
him to think innovatively and visionary. One of them was to create a vehicle
that would move independently, without any human intervention. The idea behind
this goal was broadly understood security. Vehicles of this type could, for
example, independently transport hazardous materials or move in environments
that pose a threat to human life [1].
Therefore,
the topic of autonomy, including all types of vehicles, is particularly related
to military matters. It was in this matter that the first ideas and research
appeared, and financial resources were invested. History shows that the results
of such research were often made public only after several years, and then used
for non-military purposes.
The
analysis carried out confirms the thesis that military needs were the greatest
motivation for the development of technologies enabling the creation of
vehicles that require only limited human intervention in their operation, or
even no intervention at all. We can also indicate the activities of a military
organization that directly contributed to the development of autonomous cars.
In
2003, the American government agency dealing with the development of military
technology – DARPA (Defense Advanced Research Projects Agency) announced
a competition aimed at designing an autonomous vehicle [2]. The detailed
guidelines state that the vehicle should independently cover a distance of
approximately 250 kilometers in less than 10 hours. The route was set in the
Mojave Desert, located mostly in California, but also in the states of Utah,
Nevada and Arizona, and in the southern part of the Great Basin in the USA. The
award of one million dollars was approved by the Congress of the United States
of America. The competition took place on March 13, 2004, called Grand
Challenge. Unfortunately, neither team successfully completed the designated
route from Barstow, California to Primm, Nevada. The closest to victory was the
vehicle called Sandstorm, based on the Humvee military vehicle. It was equipped
with four lidars – allowing distance determination using a laser beam,
one radar – using radio waves, two cameras, an inertial navigation system
and a GPS satellite navigation system. It traveled a relatively short distance
of approximately 12 kilometers. Ultimately, the main prize was not awarded [3].
The
next year, the winning car called Stanley, based on the Volkswagen Touareg,
completed the route in 6 hours and 54 minutes. A total of four vehicles managed
to overcome the required distance. The car was equipped with five laser
sensors, radar, camera, inertial navigation system and GPS [4].
In
the next edition, the race organizers introduced difficulties for its
participants. The race took place in simulated urban traffic at a military base
in Victorville, California. The cars had to obey California traffic rules, be
able to cope with adverse weather conditions such as fog and rain, and function
in the absence of GPS reception. The maximum stopping time was set at 10
seconds. In addition, the vehicles had to be able to perform turning and
parking maneuvers, and no collisions could occur during the competition. The
name of the competition, due to its new nature, was changed to Urban Challenge.
The winner covered a route of approximately 95 kilometers in 4 hours and 10
minutes. It was a car called Boss, based on the Chevrolet Tahoe and equipped
with four lidars, radar, camera, and GPS [5].
Looking
at the history of the development of autonomy in transport, in addition to the
activities undertaken by military-related institutions, we should not forget
about the significant contribution of civilian visionaries. One of such people
is undoubtedly Elon Musk, known as the co-founder of Tesla, a leader in
implementing the concept of autonomy in cars.
In
late 2014, the company began equipping the Model S with hardware that can
automate some steering, braking and acceleration functions. Work on
implementing the first Autopilot functions is beginning. Just a year later, the
car is equipped with a working Autopilot function, which combines adaptive
cruise control that allows driving at a specific speed and a system that allows
the vehicle to stay in the lane marked by painted lines on the road. It was not
yet a solution enabling autonomous driving, and vehicle control was still in
the hands of a human driver. However, it provided the basis for further gradual
implementation of innovative technology [6].
Analysis
of the current situation indicates that it is technically possible to introduce
autonomous vehicles on public roads, or it is close to that. However, a visible
problem arises in the legal system surrounding this technology. Deficiencies in
the relevant legal regulations hinder and often prevent this possibility.
2. ARTIFICIAL INTELLIGENCE AS A TOOL
TO SUPERVISE THE OPERATION OF AUTONOMOUS MEANS OF TRANSPORT
Regardless of the type of autonomous vehicle, the most
important part is the one that is responsible for its control without external
human participation. To make this possible, you need to use one of the tools
belonging to the group of methods called “artificial intelligence”.
People have been interested in artificial intelligence
for many years, both in the field of broadly understood entertainment –
creating films and fantasy literature on this topic, as well as in science
– both in the disciplines of engineering, technical and medical sciences,
but also in the social and humanities. The first scientific journal exclusively
devoted to artificial intelligence was “Artificial Intelligence”
(ISSN: 0004-3702), created in 1970 and still published by Elsevier. In legal
sciences, a similar journal appeared more than twenty years later – in
1992, called “Artificial Intelligence and the Law” (ISSN: 0924-8463)
and is still published by Springer [7].
The statement “artificial intelligence”
clearly refers to the concept of intelligence. Whether we are talking about
machines or people, the term can be vague. It is of interest to psychologists,
biologists, and neurobiologists. In the case of people involved in artificial
intelligence research, the concept of rationality is mainly used. It refers to
the ability to choose the best action that must be taken in order to achieve
the intended goal. The given criteria and available resources should be
considered here. Rationality is not the only component of intelligence, but it
is an important part of it [8]. It is noted that artificial intelligence at
today's level has the characteristics of rational thinking, which means not
only copying and imitating human behavior, but also fully autonomous operation.
Therefore, it can be assumed that artificial intelligence is humanocentric [9].
The concept of intelligence itself can be understood as
the ability to achieve complex goals using knowledge and skills [10]. The
following stand out [9]:
- narrow intelligence –
enables the achievement of a specific goal;
- general intelligence
– enables the achievement of any goal;
- universal intelligence
– allows you to gain general intelligence thanks to data and resources;
- superintelligence –
general intelligence exceeding the normal human level.
The term “intelligence” has been defined in
various ways by famous scientists as a skill [11]:
- abstract thinking
(according to Terman L.M.);
- learning and adapting to the
environment (according to Colvin S.S.);
- adapting to a new situation
(according to Pintner R.);
- acquiring new skills
(according to Woodrow H.);
- learning and generating
profit based on your own experiences (according to Dearborn W.F.).
It is assumed that the term “artificial
intelligence” was probably first used by John McCarthy at a conference in
Dartmouth in 1955 [12]. In his statement, he referred to the design of machines
that operate similarly to manifestations of human intelligence. With this
statement, he called the science and engineering of building intelligent
machines. In later years, he formalized the definition of artificial
intelligence as the science and engineering of creating intelligent machines,
especially intelligent computer programs. This is related to the similar task
of understanding human intelligence using computers, but artificial
intelligence does not have to be limited to biologically observable methods
[13].
In the following decades, artificial intelligence
developed in fits and starts, with periods of rapid progress interspersed with
stagnation [14]. The following years brought the emergence of forecasts
regarding the future of systems using artificial intelligence, including that
[15]:
- by the end of the 20th
century, computers will emulate human intelligence in a way that makes it
impossible to distinguish a machine from a human (Alan Turing, 1950 [16]);
- within 10 years,
intelligent machines will be created in Japan (Ministry of International Trade
and Industry, “Knowledge Information Processing Systems (KIPS)”
program, 1982 [17]);
- over a period of 20 years,
machines will gain emotions, desires, fears, love, and pride (Rodney Brooks,
2002 [18]).
It should be noted that, nowadays,
this concept is often used mainly for marketing purposes, as an added value to
a product or service, influencing consumer choices [15]. There is also a
visible tendency to abuse this nomenclature for products to which this concept
should not apply. Some of the designed systems may give the impression of being
“intelligent”, but in fact they are based on a sequence of
programming commands previously saved by the programmer, and at the same time
there is no influence of the history of the system's operation and the acquired
new data on the method of operation [19].
The most frequently cited definition
of artificial intelligence – derived directly from Turing's analyzes
– defines it as the ability of a machine to imitate or imitate human
intelligence [19].
Instead of defining what artificial intelligence is, you
can define what its purpose is. This includes activities such as reasoning,
associating, or selecting information. All this is intended to automate human
activities and intellectual activities [20].
There are two concepts of artificial intelligence:
- strong artificial
intelligence;
- weak artificial
intelligence.
Strong artificial intelligence is also called artificial
general intelligence, and human-level artificial intelligence. On the other
hand, weak artificial intelligence is also referred to as artificial narrow
intelligence [24]. One may also come across the term model, where the first one
is the connectionist model (so-called “bottom up”) and the second
one is the classic one (so-called “top down”) [22]. The difference
between them is the data that can be processed by an autonomous system.
The first of the above-mentioned types of artificial
intelligence is able to perform any task at a level no worse than that
achievable by a human [10]. It can also learn to the greatest extent possible,
and during the operation of this type of system, a mental phenomenon called
thinking or understanding may arise [23]. It is also characterized by
self-awareness and self-knowledge [24]. Furthermore, it is also defined as the
ability to think “truly”, i.e., thinking unsimulated and connected
with the awareness of one's existence [25]. From a philosophical perspective,
“thinking” itself – similarly to speaking out loud or writing
answers on paper with a pen – can be called symbolic reasoning [22].
In turn, the second type of artificial
intelligence enables the performance of only precisely specified tasks [9]. It
can be compared to a computer with intelligent behavior [26]. When considering
the difference between the mentioned types of artificial intelligence, it is indicated
that it is quantitative rather than qualitative. This means that technology
must be developed in an evolutionary manner, not in leaps and bounds [27].
According to information presented on
the European Parliament website [28], artificial intelligence is the ability of
machines to demonstrate human skills in the form of reasoning, learning,
planning and creativity. It gives the technical systems based on it the
opportunity to observe the environment, deal with what they observe, and solve
tasks to achieve the intended goal. Systems using artificial intelligence can
adjust their operation to a certain extent by analyzing the effects of
previously undertaken actions and implementing them autonomously.
3. CHALLENGES
REGARDING THE USE OF ARTIFICIAL INTELLIGENCE IN AUTONOMOUS VEHICLES
It
was noted that the risk of using artificial intelligence should not be a factor
inhibiting the development of this technology and innovative research related
to it. However, efforts should be made to anchor new technologies in human
rights and accepted moral and ethical values and principles. At the end of
2019, the UNESCO General Conference decided to develop ethical standards for
artificial intelligence. The recommendations developed were published in 2022
[29]. Their goal was to provide the basis for artificial intelligence systems
to work for the good of humans and the natural environment, while preventing
harm. They were also intended to stimulate their peaceful use. In its
recommendations, UNESCO presents an artificial intelligence system as a system
that enables the processing of data and information in a way that resembles
intelligent behavior. Its operation includes processes such as reasoning,
learning, perceiving, predicting, planning, and controlling. It is pointed out
that artificial intelligence systems:
- integrate models and
algorithms, providing the opportunity to learn and perform tasks enabling
prediction and decision-making in real and virtual environments;
- enable operation with
varying degrees of autonomy through modeling, knowledge representation, and the
use of data and correlation calculations;
- include machine learning
(deep learning and reinforcement learning), machine reasoning (planning,
scheduling, knowledge representation and inference, search and optimization);
- can be used in
cyberphysical systems (Internet of Things, robotic system, social robotics,
human-computer interface), including control, perception, processing of data
collected by sensors, operation of actuators;
- throughout their life cycle
(research, design, development, implementation, use, maintenance, operation,
trade, financing, monitoring, evaluation, validation, end-of-life, dismantling)
should consider ethical issues;
- they imply new ethical
challenges (impact on decision-making, employment and work, social
interactions, health care, education, media, access to information, digital
divide, personal data and consumer protection, environment, democracy, rule of
law, security and policing, rights of human being, freedom of expression,
privacy);
- they may strengthen
existing prejudices, forms of discrimination and stereotypes;
- are able to perform tasks
previously reserved only for living beings and even limited only to humans;
- constitute a new role in
human practices and society, and in relations with the natural environment;
- create a new context for
children and young people while growing up, developing the ability to
understand the world and themselves, critically understanding the media and the
information conveyed in them, learning to make decisions;
- in the long term, they may
pose a challenge to the human sense of experience and agency, raising concerns
related to, among others, with human self-understanding, social, cultural and
environmental interactions, autonomy, agency, value, and dignity.
The
development of legal regulations regarding artificial intelligence is
undoubtedly a difficult task. Excessive regulation may lead to the suppression
of innovation, while underregulation may lead to serious damage to citizens'
rights and loss of opportunities to shape the future of European society [30].
It
is worth noting that for the first time in history, responsibility for human
life may be entrusted to machines operating autonomously without direct human
supervision [31]. When it comes to autonomous vehicles, this raises controversy
regarding ethical responsibility for the health and life of people
participating in road traffic. It is noted that people act morally, are able to
draw conclusions and take responsibility. Machines, on the other hand, are
unable to understand the concept of morality, which results in the inability to
bear responsibility under civil and criminal law [32].
It
is indicated that the operation of artificial intelligence will be mathematical
and logical if the applicable legal system is clear and consistent [33].
The
use of artificial intelligence methods generates legal risk involving
unpredictability. It increases as the technology itself develops [34]. As a
result, damage may occur and the users or producers of a given technology may
have to bear legal liability [35]. The question arises here, who should be
liable for damage resulting from the operation of systems using artificial
intelligence. In this
matter, [12] is mentioned:
- creators of artificial
intelligence systems in the person of programmers and producers of such
systems, as well as people involved in training these systems;
- people installing
artificial intelligence software on specific devices;
- producers of devices
containing artificial intelligence systems;
- owners of artificial
intelligence devices;
- entities using systems with
artificial intelligence as part of their business activity;
- users and consumers of
artificial intelligence systems that are used for private purposes.
It
is worth pointing out here that due to the characteristics of the structure and
operation of artificial intelligence systems, proving a specific error of such
systems and indicating the cause and effect relationship is very difficult or
even impossible. This feature is reflected in the often used description of
artificial intelligence systems as “black boxes”.
It
is noted that when designing provisions on liability in the case of the use of
systems based on artificial intelligence, the interest of the person
responsible for the system should be particularly taken into account. This is
extremely important in relation to innovative systems and the resulting product
liability [36]. In the case of autonomous systems, there is an increase in the
importance of product liability regulations [37]. A situation in which only one
person is responsible for the entire risk may lead to reluctance to introduce
innovations. This in turn will have an adverse impact on the public interest
[38]. Currently, there is support for the view that the greater the risk of
using a given autonomous system, the stricter the liability mechanism should be
used [26]. This method of proceeding should be treated as an attempt to
reconcile the protection of fundamental rights with the need to develop
innovative technologies [39].
Currently,
there are no norms of international law that organize the issues of artificial
intelligence in a comprehensive and comprehensive manner. However, the authors
in [26] point out the possibility of directly applying existing legal regulations
to assess the operation of artificial intelligence. Such regulations include
regulations regarding:
- human rights;
- consumer protection;
- personal data;
- intellectual property;
- civil liability;
- competition.
At
the same time, however, it is postulated to develop a special legal regime
based on the risk principle regarding liability for damages [40].
In
2020, during the debate on civil liability for artificial intelligence, a
consensus emerged that the European Union's approach should be based on a combination
of strict liability and fault-based liability rules [41]. In the first case, a
person may be liable for the damage despite the lack of fault, which in the
second case is a necessary condition for liability to exist. Strict liability
may, for example, result from the fact of using a vehicle or from an action
that he cannot control – as is the case with an animal owner.
According
to [41], strict liability is a general presumption in most European legal
systems, while strict liability provisions constitute a narrow set of
exceptions. The document presents an analysis of selected European national
legal systems in terms of provisions regarding strict liability. Existing
differences can have a significant – even “dramatic” –
impact on those affected. An example is road accidents. In such cases, the
different levels of protection in national legal systems may be crucial for the
injured party and their relatives.
The
analysis of the issue of artificial intelligence in private international law
allows the following conclusions to be presented [42]:
- due to the complexity of
legal events related to artificial intelligence, it is impossible to define a
general statute of artificial intelligence;
- conflict rules do not serve
to legalize or outlaw, they do not decide on the legal consequences of the use
of artificial intelligence, and they do not compare or evaluate legal systems
or discriminate against foreign legal regulations related to artificial
intelligence;
- conflict rules indicate
only the applicable law;
- the lack of reference to
artificial intelligence in the regulations does not imply a statement about a
real (structural) gap – a specific loophole in the law or a technical
gap, at most an apparent (axiological) gap;
- if there is no reference to
artificial intelligence in the regulations, qualifying efforts are made to
indicate the concepts defining the conflict rule;
- the occurrence of a
situation regarding artificial intelligence does not automatically trigger the
law applicable at the seat of the authority adjudicating in a given case, but
it is possible to use foreign law, as well as the public policy clause –
i.e., exclusion of the application of foreign law indicated by the conflict
rule in the event of the possibility of an effect contrary to the fundamental
principles of public order legal state;
- when applying conflict
rules, the selection of connecting factors should provide a compromise between
legal certainty and the need to look for a law that most faithfully reflects
the analyzed relationship;
- indication of the personal
statute of artificial intelligence does not automatically define its existence
and legal personality;
- in specific cases, it is
worth determining the affiliation of artificial intelligence to a given country
based on existing ties;
- the possibility of applying
foreign law in matters related to artificial intelligence should not be
automatically rejected.
It
is important to note that the responsibility must rest with the human, not the
artificial intelligence itself. It should depend on the level of autonomy of
the system using artificial intelligence, the duration of the learning process,
or the ability to self-learn. The larger the systems are, the greater the
responsibility of the person conducting the teaching process. However, it should
be noted that the concepts of system skills resulting from the learning process
and system skills resulting from the self-learning process are not the same.
A
possible solution is the introduction of compulsory insurance, similar to the
civil liability insurance of car owners. However, such insurance cannot have
the character of currently used insurance against road accidents, which
considers human actions and erroneous decisions. In this case, it is
recommended to take into account all possible liability in the chain. A system
of mandatory insurance against potential damage may apply to manufacturers
and/or owners of systems using artificial intelligence. The created insurance
system can be additionally expanded with a special fund enabling compensation for
damage in cases not covered by insurance.
In
the case of autonomous vehicles, the introduction of a special insurance fund
is suggested [43]. It is possible that manufacturers of vehicles and the
software implemented in them may participate in this project due to the fact
that their errors cause the risk of failure and, consequently, damage. However,
the imposition of responsibility for issues related to the operation of
autonomous systems, for example for carrying out the necessary update of the
implemented software, is an open issue.
Producers,
developers, owners, or users of systems using artificial intelligence should
also consider the possibility of limiting their liability when they pay
contributions to a compensation fund or take out joint insurance to provide
compensation in the event of damage. The created fund may be general, covering
all systems using artificial intelligence, or individual for a category of
systems. The obligation to pay the premium may be one-off – for example,
when the system is introduced to sale, or periodic – throughout the
product's life cycle.
6. CONCLUSIONS
It
is indicated that systems using artificial intelligence begin to act randomly
when a situation beyond the scope of their operation occurs. This also happens
when there is no solution. The system automatically begins to act beyond its
competence, resembling a man gone mad. Even the appearance of small errors in
operation, due to their repeated occurrence, can lead to irrational and
unstoppable behavior. Therefore, it is necessary to build complex security
measures before such a situation occurs [44].
The
civil liability structure applicable to a traditional (non-autonomous) car
appears to be appropriate also in the case of autonomous vehicles. This thesis
can be put forward due to the similarity between both cases – traditional
and autonomous. In both cases, despite the lack of control over the vehicle,
the insurance holder has civil liability for the events that occurred. For
example, in the case of a traditional car, there may be a situation where its
owner did not even participate in the event causing the damage and will bear
its consequences. In such a case, similarly to the situation for autonomous
vehicles, it will not have any impact on the driver of the vehicle involved in
the road incident. Such a situation may arise, for example, when the vehicle is
driven by a co-owner or a person who is an employee of the vehicle owner [45].
The
manufacturer of the autonomous car is also pointed out as responsible for the
resulting damage [30]. However, such an approach only constitutes a change of
the responsible entity, without affecting the concept of responsibility itself
[45].
When
analyzing the operation of artificial intelligence, it should be remembered
that it is the result of intellectual work consisting in developing an
algorithm and implementing it into a computer program [46]. Additionally, it
was ordered, invented, developed, implemented, used, changed, etc. – all
by humans.
References
1.
Szczepaniak C. 2000. Motoryzacja na przełomie epok. [In English: Automotive at the turn of eras]. Warsaw: PWN Publishing House.
ISBN: 978-83-0113-228-6.
2.
Defense Advanced
Research Projects Agency. Available at: https://www.darpa.mil/.
3.
Davies A. 2017. „An Oral History of the Darpa Grand
Challenge, the Grueling Robot Race That Launched the Self-Driving Car”. Wired. Available at:
https://www.wired.com/.
4.
Thrun S. et al. 2006. „Stanley: The robot that
won the DARPA Grand Challenge”. Journal
of Field Robotics 23(9): 661-692.
5.
Urmson C. et al. 2008. „Autonomous driving in
urban environments: Boss and the Urban Challenge”. Journal of Field Robotics 25(8): 425-466.
6.
Barry K. 2021. „Big bets and broken promises: a
timeline of Tesla's self-driving aspirations”. CR Consumer Reports. Available at:
https://www.consumerreports.org/.
7.
Bączyk-Rozwadowska K. 2021.
„Odpowiedzialność cywilna za szkody wyrządzone w
związku z zastosowaniem sztucznej inteligencji w medycynie”. Przegląd Prawa Medycznego 3-4: 5-35. [In
English: „Civil liability for damage caused in connection with the use of
artificial intelligence in medicine”. Medical
Law Review].
8.
A definition of AI: main capabilities and disciplines.
Definition developed for the purpose of the AI HLEG’s deliverables.
Independent High-Level Expert Group on Artificial Intelligence set up by the
European Commission. European Union. 2019.
9.
Płocha E.A. 2019. „O pojęciu
sztucznej inteligencji i możliwościach jej zastosowania w
postępowaniu cywilnym”. Prawo
w działaniu. Vol. 40. Sprawy
Cywilne. 2019. P. 273-291. Institute of Justice. [In English: „About the
concept of artificial intelligence and the possibilities of its application in
civil proceedings”. Law in action. Vol.
40. Civil Cases].
10. Tegmark M. 2019. Życie
3.0. Człowiek w erze sztucznej inteligencji. [In English: Life 3.0. Man in the era of artificial intelligence]. Warsaw:
Prószyński Media Publishing House.
ISBN: 978-83-8169-070-6.
11. Pfeifer R., Ch. Scheier. 2001. Understanding intelligence. Massachusetts
Institute of Technology (MIT), Cambridge: The MIT Press. ISBN: 978-0262661256.
12. Wachowska A., M. Kalinowski. 2020.
„Odpowiedzialność za działania sztucznej inteligencji
– jest projekt założeń unijnej regulacji”. [In English:
„Responsibility for the activities of artificial intelligence – a
draft of the assumptions of the EU regulation”]. TKP the Law. Available at:
https://www.traple.pl/.
13. McCarthy J.
„What is artificial intelligence?”. John McCarthy's website. Available at: http://jmc.stanford.edu/.
14. Russell S., P.
Norvig. 2019. Artificial Intelligence: A
modern approach. Financial Times Prentice Hall. Harlow. ISBN:
978-0134610993.
15. Wołk K. 2018. „Czy współczesna SI
to tylko marketingowy bełkot?”. Komputer Świat. [In English:
„Is modern AI just marketing gibberish?”. Computer World]. Available at: https://www.komputerswiat.pl/.
16. Turing A. 1950
„Computing Machinery and Intelligence”. Mind LIX(236): 433-460.
17. McCorduck P. 1983.
„Introduction to the Fifth Generation”. Communications of the ACM 26(9): 629-630.
18. Brooks R.A. 2002. Flesh and Machines: How Robots Will Change
Us. NY: Pantheon Books. ISBN: 9780375420795.
19. Lai L., M. Świerczyński. 2020. Prawo sztucznej inteligencji. [In English: The law of artificial intelligence].
Warsaw: C.H. Beck Publishing House. ISBN:
978-83-8198-455-3.
20. Janowski J. 2019. Trendy
cywilizacji informacyjnej. Nowy
technototalitarny porządek świata. [In English: Trends of information civilization. New techno-totalitarian world order].
Warsaw: Wolters Kluwer. ISBN: 978-83-8160-346-1.
21. Searle J.R. 1980.
„Minds, brains, and programs”. The Behavioral and
Brain Sciences 3(3): 417-424.
22. Jankowska M. 2015. „Podmiotowość prawna
sztucznej inteligencji?”. [In English: „Legal subjectivity of artificial
intelligence?”]. P. 171-196. In: Bielska-Brodziak A. (ed.). O czym mówią prawnicy,
mówiąc o podmiotowości. [In English: What lawyers talk about when they talk about subjectivity]. Katowice: University of Silesia Publishing House. ISBN
978-83-8012-439-4.
23. Kisielewicz A. 2017. Sztuczna
inteligencja i logika: podsumowanie przedsięwzięcia naukowego. [In English: Artificial Intelligence and Logic: A Summary
of the Scientific Endeavor]. Warsaw: PWN. ISBN: 978-83-01-19492-5.
24. Rojszczak M. 2019. „Prawne aspekty systemów
sztucznej inteligencji – zarys problemu”. [In Englis: „Legal aspects
of artificial intelligence systems – outline of the problem”]. P. 1-10. In: Kinga Flaga-Gieruszyńska K., J.
Gołaczyński, D. Szostek (eds.): Sztuczna
inteligencja, blockchain, cyberbezpieczeństwo oraz dane osobowe. Zagadnienia wybrane. [In English: Artificial intelligence, blockchain, cybersecurity and personal data.
Selected issues]. Warsaw: C.H. Beck Publishing House. ISBN: 978-83-8158-596-5.
25. Kaczmarek-Templin B. 2022. „Sztuczna inteligencja
(AI) i perspektywy jej wykorzystania w postępowaniu przed sądem
cywilnym”. Studia Prawnicze. Rozprawy i Materiały 2(31): 61-78. [In English:
„Artificial intelligence (AI) and prospects for its use in civil court
proceedings”. Law Studies.
Dissertations and Materials].
26. Świerczyński M., Z. Więckowski. 2021.
Sztuczna inteligencja w prawie międzynarodowym. Rekomendacje wybranych
rozwiązań. [In English: Artificial
intelligence in international law. Recommendations of selected solutions].
Warsaw: Difin Publishing House. ISBN:
978-83-8270-013-8.
27. Księżak P., S. Wojtczak. 2020. „Prawa
Asimova, czyli science fiction jako fundament nowego prawa cywilnego”. Forum Prawnicze 4(60): 57-70. [In English:
„Asimov's Laws, or science fiction as the foundation of new civil
law”. Law Forum].
28. „Artificial
intelligence: what is it and what are its applications?”. European Parliament. Available at:
https://www.europarl.europa.eu/.
29. Recommendation on the
Ethics of Artificial Intelligence. The United Nations Educational, Scientific
and Cultural Organization UNESCO. Paris.
2022.
30. Bertolini A. 2020. Artificial Intelligence and Civil Liability.
Legal Affairs. Policy Department for Citizens' Rights and Constitutional
Affairs. Directorate-General for Internal Policies. European Union. PE 621.926.
31. Drozd W., J. Brzezińska. 2018.
„Najważniejsze problemy prawne dotyczące pojazdów
autonomicznych w perspektywie globalnej i polskiej”. Przegląd Prawniczy Uniwersytetu Warszawskiego XVII(2): 39-57. [In
English: „The most important legal problems regarding autonomous vehicles
from a global and Polish perspective”. Law Review of the University of Warsaw].
32. Hildebrandt M., J.
Gaakeer. 2013. Human law and computer
law: comparative perspectives. Ius Gentium: Comparative Perspectives on Law
and Justice. IUSGENT. Vol. 25. Heidelberg, New York, London: Springer. ISBN:
978-9400794085.
33. Goździaszek Ł. 2015. „Perspektywy
wykorzystania sztucznej inteligencji w postępowaniu sądowym”. Przegląd Sądowy 10: 46-60. [InEnglish:
„Prospects for the use of artificial intelligence in legal proceedings”.
Judicial
Review].
34. „Liability for
Artificial Intelligence and other emerging digital technologies”. Report
from the Expert Group on Liability and New Technologies – New
Technologies Formation. Justice and Consumers. European Union. 2019. ISBN: 978-92-76-12958-5.
35. Dignum V. 2019. Responsible artificial intelligence. How to
Develop and use AI in a responsible way. Springer Cham. ISBN:
978-3-030-30370-9.
36. Bożek B., M.
Jakubiec. 2017. „On the legal responsibility of autonomous
machines”. Artificial Intelligence
and Law 25: 293-304.
37. Shevchenko O. 2020.
„Connected Automated Driving: Civil Liability Regulation in the European
Union”. Teis? 114: 85-102.
Vilnius University.
38. Świerczyński M., Ł. Żarnowiec. 2019.
„Prawo właściwe dla odpowiedzialności za szkodę
spowodowaną przez wypadki drogowe z udziałem autonomicznych
pojazdów”. Zeszyty Prawnicze 19(2): 101-135. [In
English: „Law applicable to liability for damage caused by road accidents
involving autonomous vehicles”. Law Notebooks].
39. Mazur J. 2020 „Unia Europejska wobec rozwoju
sztucznej inteligencji: proponowane strategie regulacyjne a budowanie
jednolitego rynku cyfrowego”. Europejski
Przegląd Sądowy 9: 13-18. [In English: „The European Union
towards the development of artificial intelligence: proposed regulatory
strategies and building a digital single market”. European Judicial
Review].
40. Mendoza-Caminade A. 2016. „Le
droit confronté ? l'intelligence artificielle des robots : vers
l'émergence de nouveaux concepts juridiques?”. Recueil Dalloz 8: 445.
41. Evas T. 2020.
„Civil liability regime for artificial intelligence. European added value
assessment”. Study. EPRS. European Parliamentary Research Service. Brussels, ISBN: 978-92-846-7127-4.
42. Świerczyński M. 2019. „Sztuczna
inteligencja w prawie prywatnym międzynarodowym – wstępne
rozważania”. Problemy Prawa
Prywatnego Międzynarodowego 25: 27-41. [In English: „Artificial
Intelligence in Private International Law – Preliminary
Considerations”. Problems of
Private International Law].
43. Urbanik G. 2019. „Odpowiedzialność za
szkody wyrządzone przez pojazd autonomiczny w kontekście art. 446
kc”. Studia Prawnicze. Rozprawy i Materiały 2(25): 83-95. [In English:
„Liability for damage caused by an autonomous vehicle in the context of
Art. 446 of the Civil Code”. Law Studies.
Dissertations and Materials].
44. Kaku M. 2024. Wizje
czyli jak nauka zmieni świat w XXI wieku. [In English: Visions or how science will change the world in the 21st century].
Warsaw: Prószyński i S-ka Publishing House. ISBN: 9788383521985.
45. Robaczyński W. 2022.
„Odpowiedzialność za szkody wyrządzone przez pojazdy
autonomiczne”. Forum Prawnicze 1(69): 67-84. [In
English: „Liability for damage caused by autonomous vehicles”. Law Forum].
46. Robaczyński W. 2022. „Sztuczna inteligencja
– przedmiot badań czy podmiot kontrolowany”. Kontrola i audyt 6: 8-29. [In English:
„Artificial intelligence – a subject of research or a controlled
entity”. Control and audit].
Received 02.10.2023; accepted in
revised form 29.11.2023
Scientific Journal of Silesian University of Technology. Series
Transport is licensed under a Creative Commons Attribution 4.0
International License
[1]
Faculty of Transport and Aviation Engineering, The Silesian University of
Technology, Krasinskiego 8 Street, 40-019 Katowice, Poland. Email: piotr.czech@polsl.pl. ORCID:
https://orcid.org/0000-0002-0884-8765