Benefits and Ethical Vulnerabilities of Artificial Intelligence

Abstract

This research examines how artificial intelligence has evolved rapidly and dramatically influences business and personal life. The research is fundamental and focuses primarily on secondary data but also on one’s knowledge, observations, ideas, and conceptualizations. To illustrate the involvement of AI as correctly as possible, the research result is structured as follows: (1) Introduction, including the questions people usually ask and the succession of answers the research aims to do. (2) “AI overview”, where the research findings include the definition, system components diagram with the primary applicability, a brief parallel between the benefits and risks diagram in four main company activities, and some future AI trends. (3) “Ethics in AI” results in a synthesis of the ethical principles of AI, as well as the selection of fears and vulnerabilities in a diagram, followed by some ways to avoid unethical actions. Finally, the findings in (4) “The paradigm of change from Industry 4.0 to Industry 5.0”, where based on conceptualization, “the effects of AI progress in the industry are analyzed with each characteristic”. (5) Conclusion.

1. Introduction

It is incredible how technology has changed human life over the centuries and how fast it is advancing in our days!

If we go back to the end of the twentieth century, we see only in communication how mobile phones replaced landline phones. Then, by smart ones, and in most corners of the world, the Internet was already indispensable, facilitating fast communication and further development.

As of October 2023, there were 5.3 billion Internet users worldwide, which amounted to 65.7% of the global population. Of this total, 4.95 billion, or 61.4% of the world’s population, were social media users [1].

Along with other equipment and devices that have traditionally replaced them, they are nothing more than the product of so-called artificial intelligence (AI). This invention, the product of the brilliant mind of some researchers, has continuously developed very quickly, changing our daily routine and forcing us to learn new things.

The beginning was in the mid-1950s when AI pioneers (including Marvin Minsky, John McCarthy, and Herbert Simon) proposed an impossibly daring mission to recreate human intelligence in a machine, building layers of artificial neurons called neural networks [2]. The first well-publicized successes were Kasparov’s chess game with Deep Blue (1996), then Lee Sedol’s go game with DeepMind’s Alpha Go (2016), and Ke Jie’s game with Google’s improved Alpha Go (2017) [2, 3].

Other improvements and scientific discoveries followed that led to progress. The results obtained until then aroused the interest of large companies (Google and Facebook), which attracted specialists from Europe and other areas of the globe and invested enormously in research, especially after the year 2000, obtaining impressive results. These have led to the advancement of technology in the industry and in everyone’s life.

Information about AI follows one another so fast that it generates a series of questions, such as: What is AI, and what benefits does it bring? But also, what risks does it produce? What progress is expected in the future?

The answer to these questions can be found in the following successive sequences: (2) AI overview, in which, based on our references and knowledge, we briefly illustrate (2.1) The definition and principal components of AI, including some examples of AI applications in practice, to familiarize the reader with the specific terminology and make them aware of their use. Following are (2.2) some opinions regarding the benefits and risks of AI. These aspects make sense of its use, but (2.3) perspectives for future AI development will also prepare us for a holistic vision of what will come in this field.

Other questions people ask themselves are: What are the basic principles of AI ethics? Are there reasons for concern or fears regarding ethical practices and people’s safety? Are there vulnerabilities related to AI? What could be done to improve the situation? The answers can be found in (3) Ethics in AI and (3.1) General Aspects regarding Ethics, which demonstrate that ethics is not just noise but principles of conduct. In (3.2), “AI Fears and vulnerabilities” include the main fears but also concrete vulnerabilities of the system, based on the opinions from the references, which demonstrate that AI is still evolving. We are all part of this evolution, and at this stage, human errors and the reduced amount of digitized information generate mistrust and skepticism. For this reason, (3.3) the main actions that can be initiated to avoid unethical AI practices are collected. The last question is: What are the effects of AI progress on the industry? For this purpose, in (4), we present our concept called “The paradigm of change from Industry 4.0 to Industry 5.0”, accompanied by the main characteristics of the two industries, demonstrating that AI’s evolution continues. However, it is a long-term but undeniable process. It is a strategic objective that marks the whole world.

2. AI overview

2.1 Definition and components

Artificial intelligence (AI) is an advanced technology that uses real-time data to simulate human intelligence. Algorithms, computers, or robots typically run it [4]. It combines large datasets with intuitive processing algorithms by learning behavior. However, AI is a general term that covers a multitude of hardware and software based on computers.

For AI to be applied in practice, the following are necessary: equipment and/or devices, technologies, specific software and programming languages, and procedures. The main AI components are presented in Figure 1.

Figure 1.

Components of the AI system diagram [own concept].

Looking at Figure 1 from a non-IT specialist, the whole system looks fascinating! People have gotten used to some of these terms, and they seem normal. Automation, which has been used in production, has been continuously improved for over 50 years. But life seems inconceivable for those from generations Y and Z who do not use the iPhone to communicate. Anyway, most of them do not think about how AI works. AI users come into contact with the objects or services that incorporate AI and do not know or care what the respective systems’ intangible components are. However, the four quadrants of Figure 1 highlight aspects that lead to a clearer understanding of the concept of AI.

2.1.1 Equipment and devices

Many of these equipment and devices are known, and some are already in the possession of companies and individuals.

Examples [from our knowledge]: Smartphones, smart speakers, smart TVs, wearable (personal devices); washing, cooking, cleaning machines (home automation devices), autonomous vehicles, drones (transportation), security cameras, chatbots, industrial equipment, robotics (business), gaming consoles, virtual reality devices (entertainment), urban communication, traffic management (cities) and devices for health diagnosis, laboratory analyses, and assisted surgical operations (health care).

However, as technology advances, they must be replaced with more efficient ones.

2.1.2 Technology

The more frequently used and continuously improved technologies are automation, generative AI, robotics, and quantum computing, as presented in Table 1.

No.	Technology	Brief explanation of the application
1	Automation; Includes:	Monitor and control the production and delivery of products and services [4, 5].
	Deep learning	Use of artificial neural network structure.
	Machine learning	More accurate at predicting outcomes without being explicitly programmed to do so.
	Machine vision	It allows industrial equipment to make rapid decisions based on what it sees, such as defect detection, positioning and measuring parts, and identifying, sorting, and tracking products.
2	Generative AI; Models:	Learning from existing data patterns can produce different types of content, including text, images, audio, and synthetic data [6].
	Natural language processing (NLP)	Use rules to learn from existing text; allow computers to process human language as text or voice data [7].
	Artificial neural networks (ANN)	Teaches computers to process that uses nodes or neurons interconnected in a layered structure that resembles the human brain [8].
3	Robotics; Types:	It combines science, engineering, and technology to design, construct, operate, and use machines programmed to replicate, substitute, or assist humans in completing tasks of varying complexity [9].
	Autonomous mobile robot (AMRs)	Support many tasks, such as delivering items, carrying out safety and security checks, inventory automation, rescue missions, crop harvesting, and even space exploration.
	Automated guided vehicles (AGVs)	Movable robots can navigate mapped routes via wires on a floor, radio waves, cameras, magnets, or lasers. They have several applications, such as transporting heavy materials in factories and warehouses.
	Articulated robot	A typical industrial robot can range from simple two-jointed structures to complex systems of multiple interacting joints and materials.
	Cobot; (collaborative robot)	They work side-by-side with human workers. They are often compact and perform various tasks in the metal industry, automotive, electronics, laboratories, and hospitals.
	Humanoid	It resembles a generic human body form, having a specialized design, typically used for mimicking human motions and interactions.
4	Quantum computing	Utilizes quantum mechanics to solve complex problems faster than on classical computers [10].

Table 1.

AI technology (own concept; [4, 5, 6, 7, 8, 9, 10]).

As shown in Table 1, technology has evolved from automation to cobots. Many of these technologies are already applied in production or service companies, mainly because people want to make people’s work easier.

Research continues with humanoid technology and quantum computing, with a lot of passion and desire for progress through a long process that could be quantified in years or decades.

2.1.3 Software

There are four types of software: Artificial Intelligence platforms, Chatbots, Deep Learning Software, and Machine Learning Software, each with different functions, as in Table 2.

Software	Brief explanation
Artificial intelligence platforms	This will provide the platform for developing an application from scratch. It has many built-in algorithms and is easy to use with a drag-and-drop facility.
Chatbots	This software will give the effect that a human or person is doing in a conversation.
Deep Learning Software	It includes speech recognition and image recognition.
Machine Learning Software	It is the technique that will make the computer learn through data.

Table 2.

AI software [11].

The advancement of technology has created increasingly sophisticated software, which facilitates the ease and speed of performing some activities (Table 2). Chatbots and machine learning techniques are still being perfected, as they depend on the volume and quality of the data people enter into the system.

For software, programming languages like Python, R, C++, Java, Lisp, Julia, Haskell, Prolog, and Scala [12] and specific tools delivered by specialized companies are needed. For example, see [11]: Virtual assistant: Google Assistant, Amazon Alexa, Cortana; Question-answering system: IBM Watson; Machine Learning: NVIDIA, H2O AI, Google Cloud, Azure, TensorFlow; Machine Learning Chatbot: Infosys Nia; Website building: Hostinger.

2.1.4 Procedures

Specific procedures are used in AI, such as in Table 3

Procedure	Brief explanation
Algorithm	It is a set of step-by-step lists of rules for completing a specific task or solving a particular problem (e.g., sorting, insertion, dynamic programming, backtracking, etc.) [4].
Augmented intelligence	It is typically done using machine learning to analyze data and assist humans to make smarter decisions [13].
Augmented reality	It integrates digital information with the user’s environment in real time and in the real world [14].
Virtual reality (VR)	It is a computer-generated environment with natural scenes and objects, immersing users in their surroundings [15].
Artificial general intelligence (AGI)	Describes programming that can replicate the cognitive abilities of the human brain. It is a hypothetical form of AI in which a machine learns and thinks like a human [16].
Technological singularity	It is a future ruled by an artificial superintelligence that far surpasses the human brain’s ability to understand it or how it shapes our reality [17].

Table 3.

AI procedures (own concept; [4, 13, 14, 15, 16, 17]).

The first three procedures illustrated in Table 3 are already applicable and applied in many situations, but VR, AGI, and technological singularity are still in the research phase. The perspective of the applicability of these three procedures leads to most of the fears presented in (3).

Corroborating the notions from Tables 1–3, it can be concluded that AI applications depend on different concepts. In Ref. [4], four concepts are presented:

Reactive machines (similar to automation), which is limited to repetitive tasks; Limited Memory, which can analyze real-time data to make predictions and observations; Theory of Mind, which is based on real-time data, more advanced because it must be designed to understand human complexity, with individual thought patterns and past experiences that affect how they respond to certain stimuli; and Self-Aware, that will understand human emotions and feelings on a human level. [4].

Automation and limited memory are applied in different fields, such as manufacturing, energy, sales and marketing, financial services, biotechnology, media and entertainment, facilities and services, community and lifestyle, agriculture and farming, health care, education, landing and investments, and many others.

The last two concepts are still under study, but if they are accessible sometime in the future, they will build bridges across generations and mark the world differently.

According to [18], cloud migrations, generative AI, and expanded functionality will dominate enterprise resource planning (ERP) in business, such as accounting, procurement, project management, risk management, supply chain operations, and performance management. This legacy tech continues to try to adapt to meet business needs. Nevertheless, larger companies, such as IBM, Amazon, Nvidia, Facebook, Google, Microsoft, Alphabet, Tesla, and others, are twice as likely to adopt and deploy AI technologies in their business than small companies [19]. However, many companies that produce AI software, solutions, and implementation assistance are small and medium-sized, mainly in the USA, Canada, Europe, and Asia. [20]

2.2 Benefits and risks

Humanity has gained countless benefits from the development of AI. However, AI also presents a series of risks. Figure 2 shows the activities in which AI is present: production, sales and marketing, human resources, management, and business in general, and, in parallel, the main benefits it brings to companies and some of its significant risks.

Figure 2.

AI benefits and risks diagram [own concept].

Benefits: In production, it determines high accuracy for repetitive tasks [5], lowers the level of operating costs, reduces processing errors compared to the standard [5], monitors processes almost instantly for quality assurance [4], and reduces the cycle from product design until sale, obtaining a faster ROI [4]. In sales and marketing, the personalization of products and services increases [4], identifies patterns in user behavior [21], and quickly generates texts and images for promotion. In terms of human resources, saving work time as machine learning algorithms can automate tasks [21], complete task accuracy [21], and foster creativity and innovation. In management and business, it contributes to the improvement of the decision-making process [5], to the innovation of business models [4], and to the attraction of talents from anywhere in the world [4].

These benefits aim not only to replace physical work with AI algorithms but also to allow closer contact with customers and understanding of their preferences to offer them value at lower costs. AI increases the productivity and performance of the entire activity, ensuring the company’s competitive advantage in the market.

Risks: In production, human errors are possible in data integration, interoperability of systems, and the robustness of AI models [22], in data management and wrong evaluations [23], but also accidents and injuries when heavy equipment does not recognize when certain operations must be canceled [23]. Most users’ privacy policies are ignored in sales and marketing, or biased algorithms are used in offers [24]. Regarding human resources, the loss of jobs due to increased automation is possible; there is potential for bias or discrimination due to the dataset on which the AI is trained and the erosion of critical skills through the loss or ability to use them due to AI endowments [4]. In management and business, the essential risks are the bankruptcy of SMEs and some banks, missed opportunities due to the lack of investment in advanced technologies [25], and the possibility of cyber-attacks. It can be seen that the main risks are due to human errors, but since risks are based on probability, they can be managed.

The AI Act issued by the European Commission provides adequate risk assessment and mitigation systems [26]. The AI Act aims to provide AI developers and deployers with precise requirements and obligations regarding specific uses of AI. However, implementing AI is a long, challenging, and complex process that requires a lot of research, patience, application, and work to reduce or eliminate these risks in time.

2.3 Prospects for the development of AI in the future

AI development has no limits. Some of its possibilities cannot even be imagined or belong to science fiction. The world’s artificial intelligence giants, such as Microsoft Corporation, Alphabet Inc., and Apple Inc., already have AI-enabled industrial robotics development, and still, they spend billions to create those products and services.

Presumably, the very distinction between what is human intelligence and what is artificial will evaporate over time. Trends in technology include brain–machine interfaces that entirely ignore the requirements of verbal communication, robotics that give machines all the capabilities of human action, and, perhaps most excitingly, a deeper understanding of the physical basis of human intelligence. These will lead to the improvement of human and machine intelligence [4].

The main expectations foreseen by the development of AI could be the following, inspired by references [25, 27, 28]:

• Life will speed up. AI will allow large organizations to make most of the decisions much more quickly.

• End of privacy. AI systems will likely become much more knowledgeable about us than we are about ourselves.

• Formation of human-AI teams. Artificial intelligence will be used to augment human intelligence and expertise or as a partner to one or more humans working on a goal.

• The regulatory environment becomes much trickier for organizations using AI.

• AI research and experiments will lead to algorithms that make computers create.

• Several industries, such as finance, transportation, health care, and education, will be drastically affected.

The future of AI means “new normal” [29] in manufacturing systems, communication, transportation, other activities, and life. AI continues to revolutionize various industries, with an expected annual growth rate of 37.3% between 2023 and 2030 [30].

3. Ethics in AI

3.1 General aspects regarding ethics

Artificial intelligence performs according to how it is designed, developed, trained, tuned, and used. AI ethics is about establishing an ecosystem of ethical standards and guardrails throughout an AI system’s lifecycle [31].

Ethics is a set of moral principles that help us discern right from wrong. AI ethics is a multidisciplinary field that studies how to optimize AI’s beneficial impact while reducing risks and adverse outcomes [31]. Several ethical concerns relate to transparency, fairness, accountability, and the potential for AI to reinforce existing bias or discrimination. Whatever, the main three principles that came out of the Belmont Report that serve as a guide for experiment and algorithm design are Respect for Persons (recognizes the autonomy of individuals), Beneficence (healthcare ethics that is “not harm”), and Justice (fairness and equality) [32].

IBM’s pillars on AI ethics are an excellent example of practices [33]. These are explainability (transparency), fairness (equitable treatment), robustness (minimizing security risks), transparency (disclosure), and privacy (data protection). Other principles enunciated by IBM include data responsibility and environmental sustainability, inclusion, moral agency, value alignment, accountability, trust, and technology misuse. These five pillars mean AI’s design, development, guidance, and ethical use.

What is likely to happen is that, in the future, high-quality databases with historical data may be essential and unavoidable, and therefore, issues of availability, accessibility, legal compliance with legislation, and, in the case of personal data, confidentiality, consent, sensitivity, and other ethical questions [34]. The new rules issued by the European Commission aim to foster trustworthy AI in Europe and beyond by ensuring that AI systems respect fundamental rights, safety, and ethical principles and addressing risks of compelling and impactful AI models [26].

3.2 AI fears

Although AI has not demonstrated significant dangers, people still fear the expected changes. The main fears identified are loss of jobs, active learning, privacy, mind control, children’s future, and robots’ hyperintelligence (Figure 3).

Figure 3.

AI fears diagram [own concept].

3.2.1 Loss of jobs

The jobs most vulnerable to elimination are in office administration, production, transportation, banks, and others, but the reality shows that every occupation will be affected to some degree by AI. Almost 40% of employment globally is exposed to AI, which rises to 60% in advanced economies. Among workers, college-educated workers and women are more exposed to AI and more likely to reap the benefits, while solid productivity gains could boost growth and wages [29].

AI will replace jobs because AI can, within routine tasks, learn to optimize itself. And the more quantitative, the more objective the job is [28]. The lack of manual or low-skilled jobs deepens inequality. Groups and communities already vulnerable and/or marginalized are most likely to suffer the adverse effects of the disruptive change that advanced AI will inevitably produce [35].

McKinsey states that AI-related advancements may affect around 15% of the global workforce. Still, according to World Economic Forum research, AI is projected to create around 97 million new jobs, potentially countering workforce displacement concerns [30].

3.2.2 Active learning

To keep or find a job, people must learn new skills and accumulate new knowledge. Millennials (generations Y and Z) are generally passionate about AI; some are even willing to learn the new skills required in the job market. But, “much of the world is still analogue and disconnected” [36].

The World Economic Forum has produced a list of critical skills organizations will need in the years ahead. Innovation is number one on the list, followed by active learning, complex problem solving, critical thinking, lateral thinking, logic modeling, systems thinking, communications, and leadership. Unfortunately, there is insufficient time spent building a workforce with these skills [37]. These are highly specialized skills and are not easy to assimilate; that is why many fear that they will not be able to face these challenges. “People need to learn about programming like they learn a new language” [29].

3.2.3 Future of children

Many young people, especially those from Generation Z, see AI as a game and not as a means to ensure their livelihood. They use the iPhone continuously in communication and are attracted to avatars and the virtual world as a means of using time. Still, they depend materially on their parents from Generation X. Even children from Generation A have phones with which they communicate with their colleagues and spend time playing games and watching movies on Netflix to the detriment of homework given at school. “The reliance on AI can intensify isolation, potentially harm mental health, and substitute natural human relationships with virtual ones, has shown that Gen Z may be the loneliest generation ever, even before the COVID-19 pandemic” [38].

In addition, social networks can make them more depressed, anxious, and unhappy and can damage their mental health [3]. That is why Generation X’s adults fear their children’s future. “Will new technologies lead us, or are they already leading us and our children to confuse virtual communities and human connection for the real thing? Because if they do, we may lose something precious about what it means to be human” [29].

3.2.4 Loss of privacy

Society will see its ethical commitments tested by robust AI systems, especially privacy, like facial recognition, identifying people in different places, evaluating needs and health, and evaluating emotions to determine who is bored. There are few limits regarding the data collected by the technology companies that track through the Android phone the location, who the person is and who they are with, who they are voting with, as well as recording conversations through illegal access all the time by turning the speaker into a virtual microphone at laptop [3], and in other words, “surveillance is the business model of the Internet” [39]. Household appliances without a screen or keyboard will record the voice-giving commands through voice command interfaces [3]. For example, (see [3]). Google has access to personal email and to the history of searches, purchases, and credit card transactions both online and offline to offer personalized ads but also to read the content of emails by an algorithm or Netflix regarding user recognition by combining multiple data sources.

In the future, very few things will be exempt from the careful control of highly effective AI. A severe fear related to private life is the system of social credits developed by China, which could be generalized and misused [3]. The objective is to “provide benefits to the trustworthy and discipline the untrustworthy, who restrictions on employment, travel, housing, and banking transactions will punish” [3].

3.2.5 Mind control

It is about inductive bias in which the outputs are anticipated, suggesting a certain preference for specific goods or services through marketing.

AI-powered “thought decoders” will not just read thoughts; they will change them. Neural “mind-reading” decoders could spell the end of privacy. However, the full ramifications of this technology are even more problematic [40].

An immediate problem is AI hallucinations that significantly disturb user trust. As users begin to experience AI as more natural, they might quickly develop more inherent trust in them and are more surprised when that trust is betrayed. One challenge with framing these outputs as hallucinations is that it encourages anthropomorphism. Describing a false output from a language model as a hallucination anthropomorphizes the inanimate AI technology to some extent. AI systems, despite their function, are not conscious. They do not have their perception of the world. Their output manipulates the users’ perception and might be more aptly named a mirage, something the user wants to believe is not there, rather than a machine hallucination [41].

3.2.6 Hyperintelligence of robots

Perhaps the most powerful fear, as an existential threat to humans, is the idea that robots might become more intelligent than humans. Dramatic depictions of artificial intelligence as an existential threat to humans are buried deep in our collective psyche. Artificial intelligence will advance in intelligence so rapidly that it will become sentient and act beyond the control of humans—possibly in a malicious way. But, for AI to perform complex work, it must have the ability to transcribe codes on a large scale, which is now impossible, and it will take at least three or four generations for AI to become developers [42].

People fear that computers are starting to control the world or mean, heavily built robots that rebel against humans when they notice that they do not need them anymore, as in the movies (Terminator, iRobot). An example is given in Ref. [35], about the launch of ChatGPT-3 in 2022, which impressed with the ability of anyone to access the technology, and then to the launch of version 4 in 2023, which produced concerns and more than 1000 high-profile people signed a letter requesting a break of at least 6 months in the development of advanced AI. The letter suggested that advanced AI “may pose profound risks to society and humanity” and “should be planned and managed carefully and with commensurate resources”.

AI is fast becoming an alien intelligence, good at accomplishing goals but dangerous because it will not necessarily align with the moral values of its creators. And, in its most extreme version, this argument morphs into explicit anxieties about AIs enslaving or destroying the human race [43].

Not surprisingly, the media and other outlets have widely reported many of these fears. Humans struggle to recognize the implications of this new level of AI and respond proactively to something mainly beyond their control [35].

Some of these fears are partially rooted in concrete situations that happened. Still, others come from the imagination of individuals who are reluctant or feel total rejection of change, or AI will undoubtedly change their way of life and habits. On the other hand, solving these fears depends on how science is and will be used for the benefit and safety of people.

3.3 AI vulnerabilities

According to the Cambridge dictionary, vulnerable means able to be easily physically or mentally hurt, influenced, or attacked. In AI, vulnerabilities come from the interaction between man and machine. AI is a machine created and programmed by humans, so it also has some vulnerabilities that come from the lack of consideration of some risks or wrong human actions.

Although AI brings many benefits, several vulnerabilities remain, such as in Figure 4.

Figure 4.

AI vulnerabilities diagram [own concept].

3.3.1 Technological vulnerabilities

The AI system works well if the data is continuously updated. If the AI is not adequately equipped to perform data analysis and research, it needs to be fed with updated information by programmers to ensure that it keeps up with the current trends. Also, because generative AI systems consume vast volumes of data, they could be improperly governed, of questionable origin, used without consent, or contain bias. Bots can be used to create fake accounts, share deepfake videos, and spread misinformation through social media platforms. Additionally, the difficulty of distinguishing between truth and fiction can be amplified by social influencers or the AI systems themselves [44].

3.3.2 Qualitative vulnerabilities

Some applications lack authenticity and precision. While, for example, ChatGPT can conjure up comprehensive content in an instant, its messaging, branding, and relevance still lack authenticity. So are video and audio editing applications based on artificial intelligence, especially with AI-generated voice functions and avatars for presentations, with problems of precision in spoken intonations (diction, accent, pronunciation, etc.) and artificial gestures of avatars.

3.3.3 Relational vulnerabilities

AI creates addiction in relationships with AI systems and impairs human behavior with peers, such as “AI applications, including Replika, Chai, and Soulmate, that enable hundreds of thousands of ordinary people to role-play friendship and love with digital companions” [45].

AI does not understand human feelings and emotional relationships so that it can lead to isolation and more. “Avoiding human interaction can not only harm mental health, but it can create false emotional bonds and damage social relationships. Spiritual connection, unconditional compassion for the most vulnerable others, and soulful intimacy would be lost if virtual relationships were allowed to advance and replace them. human connection” [38].

3.3.4 Emotional vulnerabilities

AI is contextually incapable, and most people hate talking to bots, which do not answer most customers’ specific questions and seem limited to details. “The speed at which automation diffuses into our lives depends largely on the level of technology and people’s acceptance of robots” [46].

3.3.5 Ethical vulnerabilities

The way algorithms are created involves several ethical issues:

• Automatic bias is caused by machine learning algorithms based on statistics generated by a program based on a series of texts containing stereotypes. Therefore, it does not provide the best result and leads to discrimination, especially between genders and races. The danger lies in biased construction machines because they are trained with such data. Algorithms cannot explain how decisions are made; thus, machines only provide answers that may be wrong (e.g.,, in determining the sentence of judges or in selecting people for a job) [3].

• Many corporations and companies in the technology field are unethical and test algorithms on small, unrepresentative samples, which leads to false results [3].

• Machines do not distinguish between action and inaction, anticipated and intended effects, and direct and possible side effects. For example, chatbot applications encourage users to divulge sensitive personal health, experiences, beliefs, desires, or traumas to a chatbot, stemming from social isolation, sexual desire, need for empathy, and total neglect, which can be used for dishonest purposes by impersonators or unknown developers. Also, disclosing confidential information about the organization they work for is an absolute gold mine for any outsider or malicious person who sees chatbots as an opportunity to target state secrets [45].

3.3.6 Cybersecurity vulnerability

Although hacker detection algorithms are continuously being perfected and companies are implementing security systems and audits, AI systems are still vulnerable. An example is given in Ref. [38] regarding the latest built-in AI search engine chatbot, Sydney, which has performed destructive acts, including creating fake accounts and trolling users, hacking Web sites, deleting data, and manipulating users.

3.4 Actions that can be initiated to avoid unethical AI practices

Although it is challenging to control unethical behavior, there are still some actions that can be taken to reduce these practices, including:

1. The in-depth study of human–robot interaction and the effects of this interaction on society. “Due to the possible significant impacts of robots on society, it is important to study the interactions between humans and robots and their effects on society in general” [46].

2. Realization of algorithms with explanations regarding ethical principles (correctness, transparency in answers, etc.).

3. Develop robust AI strategies, including comprehensive data security measures, ethical guidelines, employee training programs, and contingency plans to address potential disruptions [22].

4. Modernize personnel training in organizations, especially those accessing classified information, to avoid compromise due to digital threats [45].

5. The generalization of the best practices guide in AI and the conclusion of more AI protocols for companies to follow, enabling them to avoid infringements on human rights and civil liberties [33].

6. Popularization in the media and social networks of the risks that may arise from the unethical use of AI, with examples of situations, to increase people’s awareness.

7. The development of quantum technology will protect personal data and privacy through quantum cryptography.

8. A regulation that should be issued is one in which it is universally recognized that personal data is genuinely individual.

9. Promoting non-disruptive innovative solutions through imagination and digital power that may lead to new opportunities and do not destroy industries, companies, and jobs [47].

A more human-centered approach could help us progress beyond identifying core principles and into technical execution with an appropriate balance of creativity, innovation, responsible use, and functionality [29].

The AI Act [26], the first-ever comprehensive legal framework on AI worldwide, is shaping Europe’s digital future, being a pattern for the entire world.

4. The paradigm of change from Industry 4.0 to Industry 5.0

In the twentieth century and the beginning of the twenty-first century, inventions focused on the Internet, IoT, digital platforms, big data, cloud, AI, Metaverse, robotics, and 4 and 5 G, all of which developed Industry 4.0 (around 2011), which it is characterized, in general, by cybernetic systems. “The Fourth Industrial Revolution represents a fundamental change in how we create, exchange, and distribute value. It is a technological shift merging our physical, digital, and biological worlds into one. The fast-developing technologies pushing it forward, such as artificial intelligence, genome editing, augmented reality, robotics, and 3-D printing, are promising smart solutions for intractable challenges” [29].

In the twenty-first century, profound transformations are observed in the main fields of activity due to the increasing pressure regarding competition in the global market, the reduction of pollution, digitization, innovation in the technology field, and the use of resources. Some authors [48, 49, 50] consider that it has entered a new stage of industrial development, namely Industry 5.0 (that began in the year 2020), which refers to robots and intelligent machines that work alongside people, including additional resilience and sustainability objectives.

The paradigm of change between the two industries can be found in the following areas: business strategy, technology, organizational structure, organizational culture, management, human resources, and environment (Figure 5).

Figure 5.

Paradigm of industry change [own concept].

Based on these domains of the paradigm, a brief comparison between the two industries, 4.0 and 5.0, is illustrated in Table 4.

Areas	Industry 4.0	Industry 5.0
Business strategy	It is based on mixed models (traditional and new), semi-centralized, intelligent, and personalized product/service offers.	Based on centralized modeling and simulation with digital systems, Web3 business applications, infrastructure engineering, blockchain, interactive, and hyper-personalized products/ services at lower cost offers.
Technology	Digitization in many fields; 4G development; design and 3D printing; big data, cloud; automated and early robotic processes; IoT.	Digitization and robotization in almost all fields; development of 5–6 G; virtual production (digital twins); intelligent factories; integrated metaverse and IoT; brain–computer interfaces; spatial computing; quantum computing.
Organizational structure	Networking: continuous real-time communication of man, machine, products, processes, and environment [51].	Human–robot co-working [46]. The new enterprise architecture.
Organizational culture	Different values and beliefs, new ideas about teamwork, change in the behavior of individuals, innovative corporate culture, and continuous learning based on professional collaboration.	Culture of innovation, systems thinking, control of the disruption [30], people-centric culture [49], and culture of individualism in mentality [38].
Management	A new mindset; associated attitude to digital transformation in decisions; de-centralized decision-making practices; risk aversion to change [37]	Value chain management with digital tools; AI prediction in decision making and strategy formulation; people before technology; maintaining information security awareness.
Human resources	Disappearance of some traditional jobs, new jobs in the field of informatics, and using electronic devices.	Jobs elimination and new jobs creation; skills focus on learning, innovation, and logical thinking [30]; digital mindset.
Environment	Pollution reduction technologies, waste recycling, and reducing consumption of natural resources.	Green/sustainable technology; circular economy; reduction of electronic waste
Social aspects	Discrepancies in the standard of living; social support.	Emphasizing inequality regarding well-being and standard of living.

Table 4.

Comparison between industry 4.0 and industry 5.0 (own concept; [37, 38, 46, 51]).

Industry 4.0 is all about efficiency for doing things faster and better. The dynamic is to compress time and deliver more information so workers can make better decisions.

Industry 5.0 is an evolution of that. It is inundated with artificial intelligence, robots, and automation. This increases the speed at which people are getting information. In Ref. [52], it is specified that “Industry 5.0 combines human subjectivity and intelligence with efficiency, artificial intelligence, and machine precision in industrial production, reflecting the value of humanistic care, thus realizing the evolution toward the symbiotic ecosystem.” At the core of this concept are new business scenarios aided by advanced technological themes focused on providing personalized customer experiences.

As the industry is transformed, effective policy and regulation that support businesses while also ensuring users’ rights can ultimately boost productivity—through openness to new models of collaboration and governance is best to address challenges like data privacy and mounting infrastructure demands [29].

5. Conclusion

The world is changing faster than we can imagine! Every minute, discoveries appear, and AI will be part of our lives. It will make our work easier! The great advantage is not the replacement of humans with human-like androids but the development of a multitude of computerized devices that are extremely useful, both in personal life and in organizations.

The speed of inventions and innovations in computer systems increases almost daily, and novelties are popularized in all ways. Big companies are already investing in AI, and robots work together with employees. Other companies have adapted to the new cyber requirements to maintain their competitive advantage. Using AI brings a series of benefits to businesses, especially in increasing work productivity, communication, and cost control. In addition, radical transformations are taking place in areas such as ergonomics, eco-technology, finance, and many other areas of activity. People’s appetite for free time is also changing, with young people preferring laptop games instead of spending time outdoors.

Humans give the power of AI, so big data and the algorithms used still present risks and vulnerabilities, which lead to dissatisfaction, fears, and unethical practices. These fears come from a lack of knowledge about the possibilities of AI but also because of the vulnerabilities it induces, especially the lack of ethics in some situations. Experts, researchers, and organizations must consider the ethical implications, take the necessary measures to reduce the risks, and prioritize the responsible use of generative AI, ensuring that it is accurate, safe, honest, accountable, and sustainable. Artificial intelligence systems must be secure and reliable. This requires a systematic approach to identifying, analyzing, assessing, mitigating, and monitoring risks throughout the entire lifecycle of an AI system.

However, the structure and qualification of the labor force are changing, so people are forced to acquire new skills and knowledge to face these transformations in industries.

These days, there is a growing recognition that the future of AI should not be limited to extremes; it moves toward a more nuanced form. Technology leaders recognized the temporal aspect of this and agreed that while there may be disruptions, there is time for deliberate adaptation. However, equitable access remains a challenge across all technologies, and more investment is needed to boost digital inclusion [36]. Just as oil was the natural resource for the last industrial revolution, data collected with AI will be the resource for industrial revolutions 4.0 and 5.0, and machine learning will be the new refinery for these immense databases [3].

Companies must prepare strategies for adaptability, diversification, innovation, and investments in up-to-date technology and human capital to face AI’s future challenges.

Only the main aspects resulting from the use of AI are included in those presented here. The topic becomes even more complex if one goes deeper into the details. It is also possible that this approach will become obsolete and must be updated in the short or long term. As stated in [34], the truth is, “The future of AI is full of unknowns.”