Author: Prof. Dr József Hajdú, University of Szeged, Faculty of Law and Political Sciences, Department of Labour and Social Law

This is a translated article. Original language is Hungarian. The translation was prepared on behalf of the Hungarian EPALE National Support Service.

Introduction

By using artificial intelligence (AI), it has been possible to develop solutions that used to be mere fantasy not so long ago. By now, however, they have enabled huge advances in areas that can make the world a better place for all of us, such as health, labour, finance, the environment, or agriculture.

This course focuses on the challenges of the labour market, which is increasingly faced with the consequences of the accelerating penetration of robots, digitalisation, and artificial intelligence. Today, two extreme views prevail: 1. the threat of unemployment due to technology (the robots will take away jobs); 2. technology will solve all social and environmental problems.[1] The next chapter covers two main topics: AI and Robots.

1. Artificial Intelligence (AI)

According to the European Commission’s definition, artificial intelligence (AI) refers to systems that behave intelligently, analyse their environment to achieve specific goals, and perform various actions with a certain degree of autonomy. AI-based technologies can take the form of software-only systems that operate in the virtual world (e.g. voice assistants, image analysis software, search engines, voice and facial recognition systems, etc.) or can be integrated into hardware devices (e.g. robots, autonomous vehicles, drones and smart applications related to the Internet of Things, etc.)[2].

In another approach, AI is a computer algorithm (a software) that can, within certain limits, imitate human behaviour, but at its current level of development is unable to think like humans. AI uses statistical algorithms to analyse large amounts of data (‘Big Data’) to find correlations in the data and provide results to the human user. Therefore, AI does not work without Big Data, nor without an algorithm (pre-designed action steps). 

It is important to emphasise that artificial intelligence is not a thinking machine, but a software that analyses human-generated data to find correlations in the large amount of data, and make predictions and provide answers based on these.[3]

For example, AI is used by Google Search, in personalised recommendation lists for YouTube videos, in the Facebook news feed, in traffic management, in self-driving cars, in OpenAI ChatGPT, in customer services, in healthcare, in robotic applications, etc. The list of examples is endless. 

Of particular importance is the connection with (industrial) automation, which is the automated control of machines (often robots) without human intervention, achieved through software.[4] Automation is a device or system that performs, in whole or in part, a function that was previously performed or could be performed, in whole or in part, by a human[5].

1.1 Algorithm & algorithmisation 

Algorithm is one of the most important concepts in computer science. Very simply put, if the solution to a problem is implemented according to one or several pre-designed plan(s), i.e. precisely defined steps, the solution is algorithmisation. The term refers to a chain of elementary operations, a set of rules, or a counting procedure.

Computers also work according to an algorithm that involves a series of procedures, decisions, and commands. Computers solve tasks and problems by following instructions, meaning that it performs the operation according to programmed steps, in the shortest possible way. While it is true that there are several ways to solve certain problems, the longer an algorithm is, the more room for error there is. Key elements of an algorithm: 1. problem definition; 2. planning; 3. starting point; and 4. end point.[6]

In everyday life, we constantly execute algorithms – and most of us even create them, both for ourselves and for others, such as planned routes, product assembly instructions, food recipes, etc.This involves the planning of sequences of actions and information flows, and the bigger picture is revealed only to those who understand the basics of these actions.

After careful planning, the next step is to entrust the execution of the precisely formulated algorithm to an automaton, which is the computer. This basically requires a different way of thinking, because algorithms execute instructions reasonably, control them, and sometimes correct the mistakes we make. However, automatons do not think, just execute their programming (if they are able to do so).[7]

1.2 Data (digitised)

AComputers work with data. Computers were designed to process data. To understand how computers work, we first need to understand the concepts of A) information and B) data.

A)  Information is new knowledge that is necessary for the acquirer and that can be interpreted on the basis of previous knowledge. Information is a fact that, when we learn it, we gain knowledge that we did not have before. (We can also say that the information removes an existing uncertainty.) Information that tells us something specific is also known as data. In short, information is interpreted data.[8]

B)  Data as a concept is the formalised representation of elementary knowledge, facts, concepts, or instructions, suitable for communication, representation, or processing by humans or automatic means. Data enable us to gain new knowledge as a result of data processing. Data is knowledge that can be interpreted – that is, perceived, discerned, apprehended and understood. Data are available in bulk, are managed according to specific rules, and are organised in unstructured or structured forms (e.g. tables or databases). Data can be stored in different formats (numbers, texts, images, sounds, etc.).[9]

1.3 Big Data

The Organisation for Economic Co-operation and Development (OECD) defines Big Data[10] as data sets characterised by volume (large size), velocity (continuous growth), and variety (structured and unstructured forms, such as text), and are often used by AI machines.

Big Data is ‘big’ because it involves an amount of data that is many orders of magnitude larger than was previously possible, and cannot be processed with the usual tools. Big Data represents a new phase of data analytics, and is the cornerstone of a new, data-driven economy.[11]

‘Big Data’ is not a specific tool or technology, but a concept, and a symbol of an era. In the 2010s, the number of avenues for collecting data increased exponentially. One reason for this was that the amount and quality of data traffic on the internet had increased, with more and more data being available about, for example, website visitors or digital services, and so on. From banks to the energy sector to the automotive industry, companies gained access to previously unknown quantities of data about their own activities. In fact, the data they now had access to represented a quantum leap compared to the previous eras. Above a certain amount of data, it is possible to understand a particular industrial process or digital service, or even human behaviour patterns so thoroughly that it can be used to make predictions with great accuracy. This is why Big Data has opened up a whole new era for design, medicine, software development, and even marketing.[12]

For example, Big Data can help advance medicine, invent new drugs, or understand genetic disorders. Using environmental data to better understand the planet’s ecosystem, more accurate weather forecasts can be made, among other things. All of this can help prepare for future challenges, make farming more efficient, and promote sustainable development. Shoppers may find products, services or advertisements that are better suited to their needs. This can improve the subjective well-being of consumers and help companies achieve higher profits.

However, in addition to the opportunities, Big Data brings risks as well. For example, data noise due to the huge amounts of data can lead to false conclusions. Perhaps the most pressing issue is the protection of personal data. A significant part of Big Data is composed of can be regarded as sensitive personal data. In many cases, collected data can reveal information about individuals that even they are not aware of.

Moreover, Big Data makes it possible to influence or deceive people, effectively shape their tastes, and so on. In this respect, the use and collection of data raises fundamental ethical and legal issues.[13]

1.4 Training the AI

Data is useless in itself. It needs to be analysed and interpreted in a process known as the training of the artificial intelligence. Among the many ways of doing this, the following are some of the most frequently used: 

1. machine learning;
2. the Internet of Things (IoT) and AIoT;
3. deep learning;
4. reinforcement learning.

1.4.1 Machine learning

Machine learning is a subset of artificial intelligence, where a machine tries to extract knowledge from data. The idea is to look for correlations in large amounts of data using statistical techniques. The machine is fed with prepared data, and then it is asked to come up with an algorithm to help predict the results for a new, fresh set of data. In the process, the algorithm performs enormous quantities of analysis and, once the correlations have been discovered, makes decisions and predictions about the future.

For example, one of the best-known applications is autonomous driving, where the car uses machine learning to recognise obstacles, pedestrians and other cars. Another important use is to predict or detect diseases.[14]

1.4.2 The Internet of Things (IoT) and AIoT

The Internet of Things (IoT) is essentially a set of different, uniquely identifiable electronic devices that can recognise certain essential information and communicate it to another device over an internet-based network. The ‘Internet of Things’ means nothing more than the fact that more and more devices have hardware built in to capture, transmit and receive data, enabling smart devices to communicate with each other.

The rapid rise in popularity of the term ‘IoT’ can be traced back to a very specific smart product – a wall-mounted, self-regulating thermostat, which controlled the heating. For this particular thermostat, Google paid $3.2 billion in 2014, when they bought the smart home company Nest, whose main product at the time was the self-learning thermostat. This device was one of the first in an everyday household that was no longer just ‘smart’, in the sense that it could be remotely controlled and connected to other devices, but also incorporated machine learning into its operation by observing how it was used and when the temperature was changed, and, after a while, learned how to set the right values at the right time by itself, imitating the user’s behaviour.[15]

AI solutions coupled with IoT will help it all work better and better organically, on its own, without the need to parameterise everything in advance and then make modifications manually when something changes in a person’s lifestyle. IoT digitises physical reality, making it meaningful through data, while AI is able to harness and translate Big Data into action. This is an achievement that redraws the map of the entire tech world. The main direction of development in the near future is the combination of AI and IoT, which is referred to as ‘AIoT’.

A good example is the ET City Brain, which already successfully operates in the Chinese city of Huangzhou. Based on data collected by a camera system installed in traffic lights, the system intervenes in real-time in the programming of the lights, achieving overall efficiency gains, speeding up traffic, automatically detecting accidents, parking in prohibited areas and, of course, it is also able to operate special features such as clearing the way for ambulances. This is an already functioning system that raises numerous data protection issues from a European perspective.[16]

1.4.3 Deep learning

Deep learning is a special form of machine learning based on IoT and AIoT. The fundamental difference lies in the fact that while the preparation of data and the assignment of categories and labelling rules is done manually in machine learning, it is done automatically in deep learning. An example is imaging, where a category (e.g. ‘car’) is difficult to describe manually. In machine learning, the shape of a bridge is primarily defined and assigned to a category, with additional explanation, and the algorithm will indeed identify a bridge in the image if the data match. In contrast, in machine learning, the algorithm determines the object seen, which is the bridge, with a certain probability based on the very large number of images examined.[17]

Deep learning is most commonly used in the following areas:

1. Self-driving technologies.
2. Space research, defence industry: object recognition.
3. Healthcare: disease detection using imaging techniques (e.g. MRI, X-ray, etc.).
4. Industry: in security systems that detect vehicles, equipment, etc. that pose a threat to workers.

Example: Recognising a car in deep learning. In machine learning, the shape of the car is primarily defined, assigned to a category, and the algorithm will indeed identify a car in the image if the data match. In contrast, in machine learning, the algorithm determines the object seen, which is the bridge, with a certain probability based on the very large number of images.[18]

1.4.4 Reinforcement learning

Reinforcement learning is modelled on human learning, that is, on trial and error performed in one or several, mostly dynamically changing environments. While machine learning methods work with static data, reinforcement learning can achieve results even on databases that are constantly changing. The method consists of defining certain performance indicators and then having the algorithm try to find correlations until it finds the best result. 

For example, imagine a computer game as a dynamically changing environment; a chess game, where the algorithm makes a move, the move will have a certain consequence, which the algorithm will remember, and then, after hundreds or thousands of moves, it will figure out the best solution for the situation.[19]

1.5 Some of the dangers of AI

It is undeniable that the use of AI entails real dangers. That is why there is/will be a need for legislation. Among the dangers, the following are highlighted: 

1. Many jobs are under threat from the rise of AI. (Including both lower-skilled and higher-skilled professions.)
2. AI is only as reliable as the data behind it. Therefore, AI can be wrong.
3.  The operation and training of AI is costly, and requires heavy investment, hardware, and resources. 
4. AI is not cheap, meaning that not everyone will have access to it, which, in turn, may increase social inequality.[20]

2. The robot 

The word ‘robot’ comes from the word ‘rabota’ – meaning hard work –, which featured in the literary works of Carel Čapek[21]. By robot we mean a machine capable of autonomous activity and movement, mostly with an open kinematic chain, which performs, while interacting with its environment, work that is repetitive, tedious, or dangerous for humans. As the goal is to replace humans, robots are increasingly showing more human traits, including a humanoid build and traits that resemble deliberate action. Today’s robots are complex structures based on mechatronics and IT.[22], [23] Robots can take many forms. For example, a humanoid robot resembles people, while an android is a humanoid robot that uses human movement, speech and gestures.

Even though AI and robotics are sometimes used interchangeably, they are actually different – yet related – fields. However, AI is used in many different ways in robotics. 

For example, one of the most important uses of AI in robotics is machine learning. This technique allows robots to learn and perform certain tasks by observing and imitating human actions. AI gives robots the computer vision that allows them to navigate, detect, and react appropriately. This helps them in going beyond the simple performance of repetitive tasks and in becoming true ‘cognitive co-workers’.

Another use of AI in robotics is edge computing. AI applications in robotics require real-time interpretation of the vast amounts of data collected by robot-based sensors. Therefore, data analysis is done in the vicinity of the machine instead of in the cloud. This approach gives machines real-time awareness, allowing robots to act on decisions much faster than humans can.

A robot is made up of two basic components: software (algorithms that tell robots what to do, how to react to feedback, how to optimise their performance, etc.) and hardware (including the body, the motors and the sensors.)[24]

Albeit robots are no more intelligent than humans, they do excel at certain tasks. They lack general intelligence, self-awareness and emotional intelligence. These missing qualities make you feel safe. 

2.1 Generations of robots

2.1.1 Industrial robots 

Industrial robots are robots that work in a manufacturing process. Typically, these are specialised robotic arms that perform various operations, from inserting parts to welding or painting. Industrial robots consist of three main components: 1. the arm (manipulator),[25] 2. the control box[26] and 3. a teaching device.[27],[28] The movements of industrial robots are precise and uniform, but generally not as fine as human movements, and the robot can only perform the task for which it has been programmed. A specific feature of industrial robots is that they can only work in isolation from humans (fencing, barriers, markings, etc.). This is necessary to ensure the safety of humans in the same space. An industrial robot can neither hear nor see; it can only do mechanically what it is programmed to do.[29]

An industrial robot is a robot that has a mechanical robot arm, a large range of motion – usually in three dimensions – and a separate computer-controlled actuation system, and that performs a technological operation (e.g. assembly, repair, cleaning, etc.) or an auxiliary activity (e.g. quality control, material handling, storage, etc.) related to the manufacture of a particular product.

2.1.2 Service robot – collaborative robot (cobot) – humanoid robot

1. Service robots are not used in the manufacturing industry. Service robots – and, in most cases, humanoid robots – are used as customer service assistants, receptionists, hostesses, or even hospital aides. There are also certain special applications. For example, firefighters use robots in dangerous areas to avoid risking the lives and health of human workforce.[30]
2. The Hungarian word ‘kobot’ is a simple localisation of the English ‘cobot’. Its origin can be traced back to the expression ‘collaborative robot’, which denotes the cooperation between humans and machines (robots). Even though the first, rudimentary cobot was created in the United States in 1996, a decade would pass before the first devices that were truly capable of working in the same space as humans were launched. Cobots represent flexible automation instead of so-called ‘rigid automation’. The specificity of cobots is that they are able to work directly and collectively with humans, without any distracting safety barriers or fences.[31]
3. A humanoid robot is a special service robot that imitates various human gestures and movements. Its shape is very similar to that of a human. These robots are also used to automate tasks, for cost reduction reasons. They can also replace humans in certain dangerous jobs.

Humanoid robots can only ever resemble humans to a certain extent. A robot that acts fully as a human has not yet been created – at least not that we know of.[32]

Scientists are taking the phenomenon of the ‘uncanny valley’ into account when designing humanoid robots. Basically, the more human a robot seems, the cuter it is considered to be. However, when there is a strong but not complete resemblance, people tend to see the robot as bizarre, scary, or frightening. This is the phenomenon called the ‘uncanny valley’.

That’s is why, for example, Pepper is clearly not human – it is a robot that looks very much like a human, but not very similar. At the same time, Sophia, developed by Hanson Robotics, is closer to triggering the uncanny valley sensation, since many consider the robot scary. She has the ability to mimic, has artificial skin on her face, and so on.

2.1.3 Autonomous and augmented robots

Autonomous robots work independently of human operators. These robots are generally designed to perform tasks in open environments that do not require human supervision. They are quite specific, as they use sensors to understand the environment around them, and then use decision-making structures (usually a computer) to take the most reliable next step, usually entirely based on facts and their mission.[33]

Automation vs. augmentation. These are two terms that are often wrongly confused.

1. Autonomous robot: It completely replaces human decision-making and actions with technology.

Example: assembly machines in factories. These production lines run smoothly, without any human intervention.

2. Augmentation: It supports and improves human decision-making and action through technology. Augmenting robots usually develop skills that humans already have, or replace skills that humans have lost. 

Examples include spell-check functions, such as Grammarly. Instead of replacing the need for the writer, it suggests modifications that could be made to a word or sentence structure.

Augmented reality (AR) is a technology that usually involves special headsets, glasses or projections, superimposing data or graphics on real images, using sensors and cameras to recognise the operator’s movements for feedback and control. Until now, AR has mainly been used in games, but as the technology spreads, new uses are emerging.[34]

Augmented reality (AR) is the addition of virtual objects to a person’s environment. Natural perception is enriched with three-dimensional elements. The person is constantly aware of his or her physical environment, but additional elements (texts, symbols, images, videos) appear in his or her field of vision. All this is possible with tablets, smartphones and smart glasses. 

It is necessary to distinguish augmented reality (AR) from virtual reality (VR). In the case of virtual reality, you will find yourself entirely in a virtual environment, and you can be transported into entirely different places, such as underwater or into a virtual museum. This is used by the video game industry, allowing users to experience a new ‘world’.[35]

Another virtual world that combines augmented reality and virtual reality is mixed reality (MR). In this case, we can talk about a real fusion of physical and digital reality. Here you have the possibility to move, shift, zoom in and out of virtual elements using various operations. That is, there is a real interaction with virtuality.[36]

Research on AR is still at an early stage, but the technology is expected to help extend the use of robots to more complex applications. It will help improve manufacturing quality and consistency, and increase the opportunities for collaboration between humans and robots.[37]

2.1.4 Singular robots

Technological singularity refers to a possible future point in time when, thanks to artificial intelligence, technological progress will be so accelerated and so radical that humans will no longer be able to follow, control or predict it.[38] It is also called ‘strong AI’, because these days it is very fashionable to call any machine-learning system designed for a specific task ‘AI’, in spite of the fact that actual human-level intelligence is much higher than that. It goes without saying that machine-learning systems can outperform humans by a wide margin in certain specific tasks; however, in other scenarios, they cannot even come close to the capabilities of a baby. Their biggest shortcoming is that they are not yet universal. Although they are able to learn, they can only learn in a certain area, within certain limits and on the basis of certain criteria.

Ultra-intelligent machines would, therefore, soon leave human intelligence behind, making the ultra-intelligent machine the last invention that humans would have to create. This is how I. J. Good put it back in 1965. Raymond Kurzweil, a renowned futurist, inventor and artificial intelligence researcher, author of The Age of Spiritual Machines and of The Singularity Is Near, predicts that singularity will be reached by 2045. The date seems relatively close, but Kurzweil argues that this is because of the illusion of a linear pace of perceived progress, whereas real progress is exponential. Accordingly, at the current rate, we will not experience 100 years worth of progress in the 21st century, but 20,000 years worth.

The arguments put forward by the sceptics of technological singularity can be classified into three groups:

1. They call into question the projection of exponential trends into the future.
2. They doubt the possibility of an AI equivalent to humans, and probably assume that the brain is more than an object describable by physical laws.
3. They believe that singularity is possible, but they consider it dangerous and to be avoided.[39]

Summary

We addressed two fundamental issues in this chapter. On the one hand, we presented the basic concepts of artificial intelligence (AI) in detail and in context. Within this, we focused on concepts such as the difference between the concepts of information and data, algorithmisation, Big Data, machine learning, the Internet of Things (IoT), deep learning, etc. Secondly, we looked at the main characteristics and generations of robots. In this context, the main features of industrial robots, service robots (collaborative robots or cobots), humanoid robots, as well as autonomous, augmented and singular robots were presented.

Go to the next chapter

Back to the MOOC home page

Internet resource

https://learn.microsoft.com/hu-hu/training/modules/get-started-ai-fundamentals/3-understand-computer-vision

Reference source

[1] Nem a robotok döntenek sorsunkról; https://www.jovogyara.hu/nem-a-robotok-dontenek-sorsunkrol.html

[2] A definition of Artificial Intelligence: main capabilities and scientific disciplines (2018) https://digital-strategy.ec.europa.eu/en/library/definition-artificial-…

[3] Hogyan működik a mesterséges intelligencia (AI)? Hogyan használható (2024) https://elemzeskozpont.hu/hogyan-mukodik-mesterseges-intelligencia-ai-h…

[4] Mi az ipari automatizálás? Előnyök és területei (2024) https://innomation.hu/mi-az-ipari-automatizalas-elonyok-es-teruletei/

[5] Parasuraman R., Sheridan T. B. és Wickens, C. D.: IEEE Transactions on Systems, Man, and Cybernetics – Part A: https://ieeexplore.ieee.org/document/844354 Systems and humans. 30. kötet, 3. szám, 2000, 286–297. o. 

[6] Algoritmus, https://hwellkft.hu/marketing-szotar/algoritmus

[7] Informatika oktatása / Az informatika kulcsfogalmai /B. Algoritmus; https://people.inf.elte.hu/szlavi/TAMOP-2/EgybenGeneralva/lecke7_lap3.h…

[8] Az információ és az adat. Webnode; https://tinformatika-hu.webnode.hu/tananyagok/a11-evfolyam/szakkozepsik…

[9] Orbán Anna: Adat in. Közszolgálati Online Lexikon; https://lexikon.uni-nke.hu/szocikk/adat/

[10] Gazdasági Együttműködési és Fejlesztési Szervezet: Big data: Bringing competition policy to the digital era – Background note by the Secretariat. 2016 (https://one.oecd.org/document/DAF/COMP(2016)14/en/pdf )

[11] Big Data in. Hold Lexikon; https://hold.hu/lexikon/big-data-jelentese-elemzesi-modszerek-3v/

[12] Benedek Gergő (2020) Mi az a Big Data és mire használjuk? https://lexunit.hu/blog/mi-az-a-big-data-es-mire-hasznaljuk/

[13] Big Data in. Hold Lexikon; https://hold.hu/lexikon/big-data-jelentese-elemzesi-modszerek-3v/

[14] Gépi tanulással az élvonalban (2024) https://www.magyar-elektronika.hu/tartalom/gepi-tanulassal-az-elvonalba…

[15] Benedek Gergő (2020) Mi is az az IOT? És mi az AIOT? Minden, amit a dolgok internetéről tudni kell; https://lexunit.hu/blog/iot/

[16] Benedek Gergő (2020) Mi is az az IOT? És mi az AIOT? Minden, amit a dolgok internetéről tudni kell; https://lexunit.hu/blog/iot/

[17] Hogyan működik a mesterséges intelligencia (AI)? Hogyan használható? (2024) https://elemzeskozpont.hu/hogyan-mukodik-mesterseges-intelligencia-ai-h…

[18] Gépi tanulással az élvonalban (2024) https://www.magyar-elektronika.hu/tartalom/gepi-tanulassal-az-elvonalba…

[19] Gépi tanulással az élvonalban (2024) https://www.magyar-elektronika.hu/tartalom/gepi-tanulassal-az-elvonalba…

[20] Gépi tanulással az élvonalban (2024) https://www.magyar-elektronika.hu/tartalom/gepi-tanulassal-az-elvonalba…

[21] Angolról fordítva-R.U.R. Karel Čapek cseh író 1920-as tudományos-fantasztikus darabja. "R.U.R." a Rossumovi Univerzální Roboti rövidítése. A darab világpremierje 1921. január 2-án volt Hradec Královéban; bevezette a „robot” szót az angol nyelvbe és a tudományos-fantasztikus irodalomba.

[22] https://gyires.inf.unideb.hu/KMITT/c01/ch02.html

[23] Robot https://www.techtarget.com/searchenterpriseai/definition/robot

[24] How do robots work? (2024) https://courses.minnalearn.com/en/courses/emerging-technologies/robotic…

[25] Ez a robotkarok végére rögzíthető egységeket, mint például egy megfogó, általában külön lehet az ipari robothoz csatlakoztatni.

[26] Ez biztosítja a szerkezet áramellátását és a működéshez szükséges egyéb technológiai egységeket.

[27] Ez az emberi kar és kéz működését helyettesíti. Az ipari robotoknál ez konkrétan a robotkar, illetve a rá csatlakoztatható speciális eszközök (pl. egy kéz funkciójú megfogó egység.

[28] Az ipari robotokról általában, https://www.robotvilag.hu/ipari

[29] Robotika Kiterjesztett valóság és robotok; https://www.muszaki-magazin.hu/2018/09/09/kiterjesztett-valosag-es-robo…

[30] A humanoid robot jelentése; https://netliferobotics.hu/blog/a-humanoid-robot-jelentese/

[31] Robotika Kiterjesztett valóság és robotok; https://www.muszaki-magazin.hu/2018/09/09/kiterjesztett-valosag-es-robo…

[32] A humanoid robot jelentése; https://netliferobotics.hu/blog/a-humanoid-robot-jelentese/

[33] Various Robot Generations and Robot Types (2021) https://wowknowledgeworld.medium.com/various-robot-generations-and-robo…

[34] Robotika Kiterjesztett valóság és robotok; https://www.muszaki-magazin.hu/2018/09/09/kiterjesztett-valosag-es-robo…

[35] Viride: Mi a különbség a virtuális valóság és a kiterjesztett valóság között? https://www.viride.eu/tech/mi-a-kulonbseg-a-virtualis-valosag-es-a-kite…

[36] Kiterjesztett valóság (Virtuális objektumok érzékelése a környezetben) https://l-mobile.com/hu/industrie-40/kiterjesztett-valosag-technologia/

[37] Robotika Kiterjesztett valóság és robotok; https://www.muszaki-magazin.hu/2018/09/09/kiterjesztett-valosag-es-robo…

[38] Szingularitás (2024) https://lexiq.hu

[39] Gáspár Merse Előd (2018) Mi az a technológiai szingularitás, és mikor jön már el? https://qubit.hu/2018/01/03/mi-az-a-technologiai-szingularitas-es-mikor…