View in EDS HTML Full Text PDF Full Text

Moral Considerations of Artificial Intelligence

Saved in:
Bibliographic Details
Title: Moral Considerations of Artificial Intelligence
Language: English
Authors: Sun, Fuhai, Ye, Ruixing
Source: Science & Education. Feb 2023 32(1):1-17.
Availability: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Peer Reviewed: Y
Page Count: 17
Publication Date: 2023
Document Type: Journal Articles
Reports - Descriptive
Descriptors: Moral Issues, Ethics, Artificial Intelligence, Definitions, Inclusion, Educational Technology, Technology Integration, Educational Trends
DOI: 10.1007/s11191-021-00282-3
ISSN: 0926-7220
1573-1901
Abstract: One of the ultimate problems of moral philosophy is to determine who or what is worth moral consideration or not. "Morality" is a relative concept, which changes significantly with the environment and time. This means that morality is incredibly inclusive. The emergence of AI technology has a significant impact on the understanding and distribution of "subject," which has produced a new situation in moral issues. When considering the morality of AI, moral problems must also involve moral agents and moral patients. A more inclusive moral definition is necessary for extending the scope of moral consideration to other traditionally marginalized entities. The evolving ethics redefines the center of moral consideration, effectively reduces the differences, becomes more inclusive, and includes more potential participants. But we may still need to jump out of this binary framework and solve the problem by rewriting rules. It is a huge, complex systematic project to realize moral AI in education. To be a "trustworthy" and "responsible" companion of teachers and students, educational AI must have extensive consistency with teachers and students in terms of the moral theoretical basis and expected value. The deep integration of AI and education is likely to become the development trend of education in the future.
Abstractor: As Provided
Entry Date: 2023
Accession Number: EJ1361283
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwH3rq4AHIUtP7wjlnsIZQbqAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDL7Hrntws4NZ32RuJQIBEICBm6heStpLjAIpw8-foek1CTiouQC0e9ThKfQ4cZNl2gUgl34ey303NSFkP5Gl6g2G4EdMYCHm5T3PyvmZ2fKCpMxSCbmmNj8e-e0euxOlUXfqT0HYyn4YHKDy_MAm8R47kX3atSXy445yIAQOIID9wKM0jEPJA72Rg4TBro-nbB22QWAEcJzQzvxDyBipRccVki6hhCFY84MQYFc7
Text:
  Availability: 1
  Value: <anid>AN0161191385;nmo01feb.23;2023Jan10.05:56;v2.2.500</anid> <title id="AN0161191385-1">Moral Considerations of Artificial Intelligence </title> <sbt id="AN0161191385-2">Introduction</sbt> <p>One of the ultimate problems of moral philosophy is to determine who or what is worth moral consideration or not. "Morality" is a relative concept, which changes significantly with the environment and time. This means that morality is incredibly inclusive. The emergence of AI technology has a significant impact on the understanding and distribution of "subject," which has produced a new situation in moral issues. When considering the morality of AI, moral problems must also involve moral agents and moral patients. A more inclusive moral definition is necessary for extending the scope of moral consideration to other traditionally marginalized entities. The evolving ethics redefines the center of moral consideration, effectively reduces the differences, becomes more inclusive, and includes more potential participants. But we may still need to jump out of this binary framework and solve the problem by rewriting rules. It is a huge, complex systematic project to realize moral AI in education. To be a "trustworthy" and "responsible" companion of teachers and students, educational AI must have extensive consistency with teachers and students in terms of the moral theoretical basis and expected value. The deep integration of AI and education is likely to become the development trend of education in the future.</p> <p>Human subjectivity and its moral status have always been the central issue of philosophy. "Subject" usually refers only to human beings. The concept of "agent" is more neutral than "subject." It does not need to take "human" as the premise or emphasize some social constructiveness, while Bruno Latour describes it as "through change ......to change the situation of any entity," thus expanding "agent" to all actors. Moral status means that an entity qualifies for moral consideration (Beauchamp & Frey, [<reflink idref="bib4" id="ref1">4</reflink>]). One of the ultimate problems of moral philosophy is to determine "who" is worthy of moral consideration or not. However, "morality" not only changes with the environment, but also changes significantly with time, which makes "morality" a relative concept. This shows that the concept of "morality" is flexible enough and reveals amazing inclusiveness.</p> <p>The traditional dichotomy of body and mind separates the subject from the human, thus taking (Hegel's) absolute spirit or (Descartes') spiritual substance as the starting point to explain and analyze the constitution of the human subject. Although Feuerbach and Marx have intensely criticized this explanation (Salehi & Talebzade, [<reflink idref="bib38" id="ref2">38</reflink>]), it continues in another form in modern society, namely, intelligent agent or artificial intelligence (AI). AI in software or hardware is defined as "A system that can correctly interpret external data, learn from these data, flexibly adapt and use those learnings to achieve specific goals or tasks" (Haenlein & Kaplan, [<reflink idref="bib19" id="ref3">19</reflink>]). AI simulates human intelligence through machines, especially computer systems, which means that computers can simulate intelligent behavior with minimum human intervention. The key is that the research and application of AI always revolve around "intelligence," that is, to copy human intelligent behavior. With the rapid development of science and technology, the adaptive ability and learning ability of AI have been continuously enhanced, and its autonomous ability has begun to turn to internal development. This growing autonomous intelligence makes AI expected to undertake the responsibilities and obligations that belong to human beings.</p> <p>The emergence of AI has made a significant impact on human understanding of the "subject." AI that can learn independently based on neural networks and genetic algorithms has produced new moral issues. In principle, designers, manufacturers, and users can no longer predict and fully control AI's behavior, and thus cannot bear moral responsibility. The traditional attribution of responsibility conflicts with the moral framework. In other words, no matter how primitive the autonomous behavior of AI is, the previous technical definition of subscribing machine behavior to human subjects is no longer suitable for designing and exhibiting autonomous operating mechanisms. Therefore, it needs to be clear whether AI needs to be responsible and liable for its behavior? Is it possible for AI to be regarded as a legitimate moral agent? And what does it mean for humans to extend the moral agent to AI?</p> <p>Moral considerations in intelligent systems design are becoming an important field of AI research. If AI can act morally, it must be endowed with a mechanism for moral decision-making and dealing with moral conflicts. For instance, self-driving technology and autonomous weapons will directly affect human safety. But the power of life or death is not the only way AI can affect human beings, its potential impact goes beyond physical damage. Just as discussed by Sharkey ([<reflink idref="bib43" id="ref4">43</reflink>]), robots deployed as teachers or teacher assistants in the classroom may need to decide which behaviors of children are acceptable or punishable.</p> <p>Let us take AI teacher as an example. AI can diagnose students' learning status and design learning paths for them. Its task also includes reminding students to learn, but sometimes students refuse AI teachers' suggestions. Then, AI needs to make a moral balance between its own obligations and students' freedom. But even if AI can act in a moral way, what moral claims human beings hold on AI are still the first priority to be solved. However, on the problem of what kind of moral stance might be given to non-human entities such as animals, machines, or AI, the most common method in history is to compare it with human beings themselves. In other words, it depends on whether there are fundamental differences between humankind and other entities, and whether an entity must possess certain attributes to be regarded as a moral agent or obtain some rights. This method is based on the assumption that two different types of entities can be easily determined. But human characteristics are often difficult to define precisely, and even some characteristics are still controversial. Moreover, when the world was divided into various categories and hierarchies by humankind, the importance of hypothesis is often ignored intentionally or unintentionally.</p> <p>Another approach usually does not talk about which attributes humans have and whether they are unique, but discussing what assumptions are assumed in interpreting humans and whether these assumptions are unique to humans. For example, concrete individuals and their groups, classes, nationalities, and countries in the past, present, and future all show their different explicit characteristics. So how to understand what a human being is? The typical simple assumption is that human exists as a species, which must have something common between people, i.e., universal human nature. Recognizing this universality, many theories put forward various human nature assumptions from different angles, thus having multiple moral propositions. For instance, MacIntyre ([<reflink idref="bib28" id="ref5">28</reflink>]) believes that human identity is the identity of animal nature. Any moral philosophy theory about goodness, rules, or virtue that ignores human animal nature cannot explain the biological existence of human beings, because we cannot deeply grasp the importance of virtue in human life without understanding the initial animal state of human beings. Can animals use this hypothesis? Singer ([<reflink idref="bib44" id="ref6">44</reflink>]) believes that we should extend the basic moral principles of equality that apply to all human beings to the other species. This is the first basic proposition about the moral status of animals based on Bentham's thought, "all animals with feelings, whether human or non-human, should be regarded as morally equal, so equal concern should be given to their same interests."</p> <p>Although there are many differences between human beings and others, these differences are not enough to deny non-human moral considerations. Currently, humankind is facing another big challenge, that is, the moral consideration of AI. It is necessary to explore the moral attribute of AI from the development of human's moral claims for themselves and animals. The analysis of the essential attributes of tools, technologies, and autonomous technologies will contribute to the discussion of the features of AI technology and what moral propositions should be held. The way humans respond to this challenge will profoundly impact how we understand ourselves, our position in the world and relationships with other entities.</p> <hd id="AN0161191385-3">Moral Consideration of Human, Animal, and AI</hd> <p>The moral consideration of human beings and others is a kind of social practice after all. It is unnecessary to treat different entities exactly the same way, or different entities must have the same rights. The key is whether we should consider allowing the same treatment or different rights. For example, the same interests need the same attention, and individual differences also need different treatment. Let us start with the moral considerations of human and animals to see what kind of moral proposition we should hold for AI. The significance of moral considerations for animals depends on the moral claims for other species or entities in any particular situation. Answering this question will enable us to understand better human nature and the appropriate scope of our moral considerations.</p> <p>It is complex and difficult to define "Human." In the Aristotelian tradition, "human" is defined as "Rational Animal." Immanuel Kant believes that "human beings are natural beings with rationality," and morality is completely transcendental in human rationality. And Ikäheimo and Arto ([<reflink idref="bib24" id="ref7">24</reflink>]) answer to this question from moral status is very representative. First of all, human is a rational creature, and therefore get a special moral status; secondly, it is obvious that only rational creatures can claim, recognize, or respect moral status. So, on the issue of how much human beings have uniqueness and enjoy the moral status, it is traditionally believed that this depends on whether human beings have some abilities or characteristics that other entities lack.</p> <p>In human history, many cultural groups have stratified the values of human beings according to their gender, religion, race, nationality, or even social class. During those periods, the attitude to human rights has changed greatly, and in moral issues. From the development process of moral philosophy, it is always breaking through its own limitations. Entities previously excluded, such as foreigners, women, and animals, are gradually being considered. Descartes ([<reflink idref="bib14" id="ref8">14</reflink>]) connected animals with machines and introduced a strong concept: biological machines or animal machines. Descartes never doubted that animals are machines, and they are interchangeable in essence. This viewpoint does not recognize that animals are the object of moral concern, but believes that they are just tools that human beings use more or less effectively, and human beings are the only subject. Immanuel Kant also believes that only rational people can become the object of moral relations. Animals are classified as irrational beings for lack of humanity, so there is only tool value. But Kant also believed that human has indirect obligations to animals (O'Hagan, [<reflink idref="bib36" id="ref9">36</reflink>]).</p> <p>Kant's ethics advocates rationality and human value, but denies the moral consideration of irrational person or non-human. It was not until modern times that animals were regarded as the object of moral consideration. As Shapiro ([<reflink idref="bib42" id="ref10">42</reflink>]) said, there are many ways to define the moral agent; the choice of definition is the key factor whether the moral subject is limited to human beings. The definition itself is a decision with moral consequences, and who should have moral status or who should be excluded have been predetermined (Misselhorn, [<reflink idref="bib30" id="ref11">30</reflink>]).</p> <p>According to animal behaviorist Frans de Waal ([<reflink idref="bib47" id="ref12">47</reflink>]), evolution has produced the preconditions of morality, such as the development and compliance with social norms, compassion, mutual help and justice, and conflict resolution mechanisms. Therefore, we should not simply assume that moral characteristics (such as courage, compassion, loyalty) exist only in humans and not in animals at all. Some animals not only help each other, but also sacrifice individuals to help others. This means that animals not just have the ability to perceive good and bad, and have the ability to weigh choices and make decisions independently, which can be said to have a certain subjectivity. According to Kant's principle, animals have freedom to their bodies, so they must have some rights. Modern scientific research also shows that animals have different characteristics, which are generally more than inorganic bodies, but some features are less than human beings (Calverley, [<reflink idref="bib10" id="ref13">10</reflink>]). Nevertheless, some characteristics (natural capabilities such as needs, emotions, desires, perception, and behavior) are the effective support for moral consideration of animals to a certain extent. Animals interact with the environment from the perspective of organizing their needs, interests, and physical abilities. This does not require animals to have abstract characteristics such as intelligence or rationality. As long as these needs and desires are valuable to the subject, they should be given moral attention and concern.</p> <p>AI seems not to be in such moral considerations at the beginning, but its increasingly advanced independent brain and powerful machine learning ability not only play an irreplaceable role in society, but also surpass human beings in many fields. The development of AI will eventually evolve into artificial general intelligence (AGI) or artificial super intelligence (ASI). Maurice ([<reflink idref="bib32" id="ref14">32</reflink>]) stated that intelligence lays the foundation and presupposes a more basic way of coping shared by humans and animals. Therefore, artificial intelligence, animal intelligence, and human intelligence are on the same continuum. For easier understanding, simpler animals can be seen as a bottom-up model for building intelligence. But the difference is that AI has both disembodied intelligence and embodied intelligence. More specifically, AI is a disembodied cognition or mind, which is easy to understand. However, AI can also physically exist in the form of the embodied autonomous system. It is usually a combination of the following characteristics: autarchy, independence of human control, interaction with the environment, learning, and mobility (Burri & Trusilo, [<reflink idref="bib7" id="ref15">7</reflink>]).</p> <p>Similar to the moral proposition of animals, whether based on empathy or social values, we cannot refuse more and more intelligent AI. However, we must consider the moral of AI and its responsibilities, obligations, and rights. As in the previous example, robots in the classroom need to make real-time explanations and decisions. Their decisions with moral consequences must comply with the moral norms and values of human teachers. If there is no moral ability, how can robots make appropriate decisions on praising or restricting students' activities? Therefore, AI has a moral dimension here, which not only plays an active role, but also a moral bearer. The higher autonomy of AI, the more moral rules are needed.</p> <hd id="AN0161191385-4">Moral Consideration of Tool, Technology, and Autonomy</hd> <p>Tool use is considered to be a sign of intelligent behavior. "Tool" originated from ancient English "Tawian," refers to any form of things that can directly or contribute to the execution of tasks (Baber, [<reflink idref="bib2" id="ref16">2</reflink>]). The user manipulates tools in a way that represents the user's extension and affects the environment. According to Butler ([<reflink idref="bib8" id="ref17">8</reflink>]), tools can only become tools in the process of use, and the essence of tools lies. Tool-using mediates the interaction between the user (human, animal, etc.) and the environment, which is usually considered as a goal-oriented behavior, representing the user's intention. Therefore, the user must be responsible for using the tool, and the tool itself cannot be the subject of responsibility.</p> <p>Technology is usually regarded as a tool or means, rooted in Western philosophy tradition. Plato believed that written language is a supplement to human voice and memory, Socrates also expressed this point clearly, that writing is just a tool and had no meaning (Naas, [<reflink idref="bib35" id="ref18">35</reflink>]). The tool theory is even applicable to modern technology. For robots and AI, Bryson ([<reflink idref="bib6" id="ref19">6</reflink>]) thinks that there is no difference from any other technical tools. They are tools made and used by human for specific purposes, so they are only "extensions of users." McLuhan ([<reflink idref="bib31" id="ref20">31</reflink>]) described all technologies as media, which are actually a transmitting means, and all media are extensions of human. But McLuhan also admitted that with the advent of computers, technology is no longer just an extension of human perception and motor functions, but an extension of cognition.</p> <p>There is an essential difference between technology and tools. In Heidegger's view, technology is neither art nor skill, but a way of cognition. Therefore, in terms of meaning, tools are the product of technology. What determines technology does not lie in manufacturing and manipulation, nor the use of tools, but the interaction between the regulation of the world and the intention of the subject.</p> <p>Despite many discussions about technical morality, the moral subject is still assigned to human designers or users. As Heidegger said, the tool definition of technology is based on the assumption of human subject status, while technology is often placed in the position of "moral neutrality." Channell ([<reflink idref="bib12" id="ref21">12</reflink>]) considered that its external factors determine the mechanical object's moral value and depend on its usefulness to human beings. Searle et al. ([<reflink idref="bib41" id="ref22">41</reflink>]) even thinks that computers' importance, like any new technology, has been greatly exaggerated. Computers are just helpful tools. Therefore, the problem of subject, especially moral subject, lies in the subject of its design, use, or implementation. According to those analyses, no matter nails, clocks, or computers, technology itself does not participate in the big issues of truth, morality, or esthetics.</p> <p>But if technology is only a purposeful tool or means, why control technology? Cybernetic expert Norbert Wiener pointed out that technology can help human beings become better and create a more just society, but we must control technology to do this (Cavalier, [<reflink idref="bib11" id="ref23">11</reflink>]).</p> <p>Technology embodies the relationship between human and the world. It is not only the way of human existence, but also the way of human self-construction and world construction. In the interactive relationship between human beings and the objective world, technology not just affects human existence in new ways, itself is under the tension of morality all the time. This definitely shows that technology can adjust the subject, and technology has the moral attribute, which means that technology must take a moral stance. The moral problems caused by technology are not just resulted from its abuse. As mentioned earlier, the essence of technology determines that it is also an inevitable result of "intentional structure" and "value embedding." Due to the certain incomplete consequences of application and the possible deviation between technology path and behavior path, it is necessary to intervene in technology design. Once the expected moderating effect is integrated into technology, it will lead to a clear behavior impact: technology moralization.</p> <p>The emergence of autonomous technology has brought greater challenges. Autonomy comes from the Greek word "auto" (self) and "nomos" (rules or laws) (Scott, [<reflink idref="bib40" id="ref24">40</reflink>]). Autonomy is a complex concept, which cannot be simply understood as independence or self-control. It is limited by but not entirely determined by the domination of external laws (heteronomy). In Kant's metaphysics, autonomy is the basic condition for realizing free will, that is, the ability of will to follow its moral rules (Winner, [<reflink idref="bib51" id="ref25">51</reflink>]). Autonomy is not isolated, nor is it antagonistic to heteronomy. Heteronomy includes the principles or norms that the subject obtains from external resources to govern the system's behavior. In fact, it is both constraint and resource. The existence of heteronomy will not make the subject lose autonomy. According to Ellenberg, a subject's real autonomy is not to create its own norms for self-satisfaction, but to follow the existing norms. In other words, full autonomy requires heteronomy (Steiner & Stewart, [<reflink idref="bib45" id="ref26">45</reflink>]). Based on this, autonomous technology involves the expected relationship between agent and patient. The agent's behavior responds to external stimulation or restraint, and a behavior involving responsibility, just because of heteronomy. Therefore, the interaction involving at least two entities constitutes the basis of social and moral relations.</p> <p>So, autonomous technology is not only a tool used by human beings, but may gain the subject position in another way. As Marx ([<reflink idref="bib29" id="ref27">29</reflink>]) said, a machine is a tool that performs the same operations as a worker. By this logic, autonomous machines can separate tools from goals, and independently select and achieve goals, which is both autonomous and intentional. Therefore, this autonomous mechanism not only replaces workers' hand tools, but also replaces workers themselves. So, autonomous technology itself should and must be a moral agent.</p> <p>Arthur ([<reflink idref="bib1" id="ref28">1</reflink>]) discussed that "with the development of gene research and nanotechnology, biology is becoming technology. At the same time, from the perspective of technological evolution, technology is also becoming biology." Although there is a bit exaggerated, it raises a very interesting question. With the development of machine learning, will AI technologies break away from human beings and begin to evolve them self?</p> <hd id="AN0161191385-5">Artificial Intelligence, Moral Agent or Moral Patient?</hd> <p>Morality has different traditions and definitions in human history. Morality describes a social interaction that involves at least two entities. Floridi ([<reflink idref="bib16" id="ref29">16</reflink>]) believes that any behavior, whether morally responsible or not, is a binary relationship between agents and patients when reduced to the smallest logical structure. Here, the concepts of agent and patient come from subject and object in philosophy, respectively. The subject is the undertaker and implementer of cognitive and practical activities, while the object is cognition and practical activities. Obviously, subject and object are not just names, while announce a special relation between subject and object, which is the relation of cognition, practice, and even value. In typical cognitive activities, the distinction between subject and object is used to describe their respective positions. Subject puts the things on the "opposite" and observes or analyzes it rationally, so as to determine the facts about object or the mechanism behind those facts. Actually, the subject's cognitive attitude towards the object is a de-relational and de-emotional attitude. This is what Kant defined "rationality."</p> <p>The most successful and exemplary representative of this dualistic cognitive activity is science, especially natural science. The boom of modern science and its great changes all reflect the huge advantages of the dualism. Of course, the disadvantages of dualism are also obvious. It separates the natural state of human social life and deeply obscures the natural face of life. Heidegger's concepts of "zuhandenheit" and "vorhandenheit" deeply uncovered this (Lindemann, [<reflink idref="bib27" id="ref30">27</reflink>]). Therefore, people tend to ignore the moral relevance between entities and the whole involved, so they cannot see the impact of social relations and society on moral considerations.</p> <p>When considering the morality of AI, moral issues must also involve the agent and patient. Moral agents may include all types of entities from the source of moral behavior, and moral patients may include all types of entities of recipients of moral behavior. According to Floridi and Sanders ([<reflink idref="bib17" id="ref31">17</reflink>]) research, there are five possible logical relations between moral agents and moral patients. From the perspective of the entities that moral agents and moral patients may contain, three kinds of relations are impossible to establish in the real world, including that agents and patients are totally unrelated (Fig. 1). For example, human beings are both moral agents and moral patients, so completely irrelevant relations are impossible. The other two relations are that moral patients are subordinate to moral agents (Fig. 2) and that the two only intersect (Fig. 3). If the latter two relations are valid in reality, there is at least one moral agent that is not a moral patient exists in real life. Then, the moral agent must be a supernatural entity that affects the whole world without its influence.</p> <p>Graph: Fig. 1 The two unrelated (an impossible relation)</p> <p>Graph: Fig. 2 Moral patients is subordinate to moral agents (an impossible relation)</p> <p>Graph: Fig. 3 The two intersect (an impossible relation)</p> <p>Two other relations are possible in real social life. The first is that moral agents and moral patients are equal (Fig. 4). All entities with moral agent qualification have the qualification of moral patient, and vice versa. This kind of relation is very intuitive, because that is what humans are. It is the most popular view of "human center" in the history of ethics, which has been recognized by many ethicists such as Kant. The second is that the agents may be an appropriate subset of the patients (Fig. 5). All entities eligible as moral agents are also entitled to be moral patients, but otherwise not so. Therefore, even many entities (animals, etc.) excluded from the moral agents in principle are eligible to be moral patients. This is a view from agent orientation to patient orientation.</p> <p>Graph: Fig. 4 The two equals (a possible relation)</p> <p>Graph: Fig. 5 Moral agents are subordinate to moral patients (a possible relation)</p> <p>For human beings, adults are usually regarded as full moral agents, who need to be responsible for their actions and enjoy rights. Human children usually have the right to be protected and bear limited responsibilities before adults, while newborn infants have moral rights but no ability to bear any responsibilities. AI is likely to begin with the necessary independent decision-making and behavioral responsibility and gradually have limited rights. Moral initiative is a continuum from non-morality to full autonomous morality. Some studies have put forward the moral level of AI from low to high according to autonomy and moral sensitivity, which includes operational, functional, and full moral behavior (Wallach & Allen, [<reflink idref="bib48" id="ref32">48</reflink>]). Similar to human children, AI is also developing towards (or even beyond) human capabilities. However, there is no pre-existing correct answer to the question that whether AI is or may be considered as a moral agent or moral patient. This needs further discussed.</p> <hd id="AN0161191385-6">Consideration of AI as Moral Agent</hd> <p>There are many ways to define moral agent, and the choice of definition is the key factor of whether moral agent is proved to be limited to human beings. As mentioned earlier, any definition of the moral agent itself has a decisive moral consequence, which predetermined who has moral status. Steve Torrance ([<reflink idref="bib46" id="ref33">46</reflink>]) holds that moral agent refers to any agent responsible for moral behavior. To some extent, the definition changes the prior hypothesis that the entity's moral status is limited to the human subject. Therefore, if the lower standard is accepted, there is no doubt that entities such as animals or AI are worthy of moral consideration.</p> <p>Bostrom and Yudkowsky ([<reflink idref="bib5" id="ref34">5</reflink>]) points out, if two entities have the same function and consciousness experience, but only different in the ways of realization or formation, they will be considered to have the same moral status. Thus, similar to the human subject, a non-human entity will have the subject status if it acts based on internal motivation. If it is moral reasons, it will have the status of moral agent.</p> <p>AI may play roles as moral agents in three aspects: explicit moral agents, implicit moral agents, and full moral agents (Moor, [<reflink idref="bib34" id="ref35">34</reflink>]). Explicit moral agents have clear moral norms, so AI can recognize and deal with moral information in various situations and respond to the environment in a moral way. Therefore, it also means that explicit moral agents must involve explicit moral reasoning, clearly expressing moral rules and guiding behavior. The design of implicit moral agents is generally implied in the process or object. Whether people realize it or not, the design will introduce the moral norms and values held by human beings (designers and stakeholders), and the system will make decisions and behaviors with moral impact in specific situations. In the implicit level, both the moral agents and the moral patients tend to be diversified, and the moral norms also change from the explicit transcendence to the pluralistic complexity. For the full moral agent, as far as human beings are concerned, is both a moral agent and a moral patient, adults are the typical representative of the full moral agent. Of course, not all entities are so, including humans. For example, children are already moral patients before they are regarded as legal moral subjects. AI, on the contrary, will be the responsible moral agent before obtaining legal rights. The closer AI is to the full moral agent, the more urgent the problems that human beings need to face.</p> <p>Just like human beings, there are two design orientations of moral agent: moral universality and moral particularity. Moral universality claims that the right decision is determined by the moral rules applied in the situation (rule-oriented), and AI based on deontology fall into this. However, the moral particularity emphasizes the situational elements (situation-oriented); Aristotle's virtue ethics has the brand of moral particularity. Virtue ethics hold that human beings learn moral behavior through practice, and virtue is not always reflected in the subject's behavior, but triggered under appropriate conditions. For example, being brave does not mean always doing brave things, but acting bravely when necessary. Howard and Muntean ([<reflink idref="bib23" id="ref36">23</reflink>]) developed a virtue AI model with the goal of replicating "the moral behavior of human actors," which emphasizes pattern recognition and situational learning (i.e., combining machine learning technology with virtue ethics) through moral example. The purpose is to establish a group of agent characters supporting virtue behavior (or similar behavior). Ideal virtue AI agent can adapt to the new situation and make moral decisions without rules. The current level of AI may not be able to meet the relatively high moral standards set by traditional ethics such as deontology, consequentialism, or virtue ethics. However, according to the degree of AI's participation in society and its ability to act morally, moral considerations need to be taken as a whole, comprehending the movement of the moral system and together of the social milieu depends on to save mutually, each other influence of dynamic state relation, which provide a new science angle of view.</p> <p>There are still different definitions and standards for moral agent, which reflects the attempt to extend moral considerations to non-human beings. AI obviously cannot meet the higher standard as a moral agent, but the higher standard is still controversial for human beings. Therefore, the machine problem about whether and when AI becomes the moral agent, it is, essentially, the question of whether human beings should expand moral standards and include AI. If moral standards are extended to AI, it means that AI needs to take responsibility, and that human beings should be responsible for AI.</p> <hd id="AN0161191385-7">Consideration of AI as Moral Patient</hd> <p>Traditional ethics only focuses on human character (virtue ethics) or human behavior and result (deontology, consequentialism), so it is inevitably human-centered. "Moral patient" is a dialectical concept, which is the opposite of moral agent (Hajdin, [<reflink idref="bib20" id="ref37">20</reflink>]). Therefore, the emergence of moral patient is mainly to deal with exclusivity, focusing on whether animals, machines, robots, and other entities can be included in the moral community, and explore what extent other entities can enjoy moral rights. Calarco ([<reflink idref="bib9" id="ref38">9</reflink>]) held that from the dominant ethics, Kantism and caring ethics had left no place for animals. Animal ethics can only develop its stance through the same exclusion, committed to establishing a more open and broader ethic. Singer believes that utilitarianism can be used as the moral claim of non-human animals, while Tom Regan refused singer's utilitarianism (Rowlands, [<reflink idref="bib37" id="ref39">37</reflink>]), believing that many non-human animals have spiritual rights and have the same inherent value as human beings. Floridi ([<reflink idref="bib15" id="ref40">15</reflink>]) put forward a new macro ecological ethics centered on patients, which is a more universal form of ethics, not only shifts attention to moral patients, but also expands the qualification.</p> <p>Generally, the rights granting to one group (such as animals) will limit the rights of another group (usually humans). It is a gradual process for animals as moral patients to obtain some rights, which prompted a paradigm shift in moral philosophy's basic structure and procedure. On the one hand, through questioning the uncensored privilege given to human beings, animal ethics challenged the human-centered moral tradition, and challenged the established moral system of humankind too. On the other hand, this transformation means that the once sealed field of ethics is opening to non-human. That is to say, not only other "people" are considered by morality, but also all kinds of marginalized entities outside the door of moral community.</p> <p>Although animal ethics is an unprecedented innovation, it still has potential prejudice. For example, Reagan continued to exclude certain animals as machines, which were undoubtedly outside the moral consideration. There has been great progress and innovation to improve the inherent moral problems by introducing another center or standard, but animal ethics unintentionally provide a universal moral consideration, and this substitution does not fundamentally change the game rules.</p> <p>Moral considerations in AI are also exclusive, such as Torrance's ([<reflink idref="bib46" id="ref41">46</reflink>]) "organic view of ethical status." This dichotomy distinguishes two types of entities; one is organisms with perception, so they have the natural legal status of moral consideration; the other is non-organism, such as machine, which is not covered by moral consideration. As Torrance recognized, this division strategy explicitly excludes entities such as machines from the qualification of moral considerations. However, AI has changed people's cognition of non-organisms, and the understanding of difference between intelligent mechanism and organisms will be refreshed.</p> <p>From the same viewpoint, the protected rights are also applicable for at least some types of AI. Usually, there are three factors for an entity to obtain the right of protection. The first is increasing happiness and reducing pain, which is Bentham's "Felicific Calculus" (Baujard, [<reflink idref="bib3" id="ref42">3</reflink>]), including humans and unlimited entities. Thus, the morality of AI maybe understood as a series of reasoning rules, which logically lead any rational existence to the same moral decision, and stimulate the utilitarian calculus of summing up happiness and pain.</p> <p>Whether an entity should have rights for the possibility of suffering seems to imply the assumption that the entity knows or realizes itself suffering. Consciousness is not unique to human beings and has objective characteristics that can be defined and observed. AI has begun to have varying degrees of sensory abilities; there is certainly the possibility of developing AI with consciousness. Although AI technology can implement similar mechanisms, the question remains whether an AI system is entitled to partial rights if it gains awareness? However, with or without consciousness, AI has been or will participate in moral decision and behavior, and its legitimacy needs to be established.</p> <p>Empathy is also an important reason to protect the others. For human society, understanding other people's feelings is the foundation of establishing common values. Kant ([<reflink idref="bib25" id="ref43">25</reflink>]) even extended this to "responsibility for non-human." The emotional response of empathy enables humans to cooperate with others, extending beyond the family, race, or country. Because of the similar characteristics, human beings have the same emotional response to animals, while AI can enhance human empathy when it continues to learn and improve its behavior. So, the more intelligent AI evolves, the more likely it is to exhibit this emotional response. Of course, the differences between animals, AI, and humans determine different rights, but the granting of rights to entities depends not only on the entity itself, but also on human attitudes. So, whether AI is endowed with rights or not, human beings need to think beyond the existing concepts and frameworks.</p> <p>Finally, in order to obtain the rights of protection, the entities must have value (inherent value or external value) and be able to benefit from its rights. The entities without value cannot be represented. Inherent value is a view of natural rights. Whatever it is, human, animals, or AI, has its intrinsic value. As far as its essence and definition are concerned, inherent value has nothing to do with the happiness and pain that the subject has or feels. It regards the existence of individual entity as its inherent value. Therefore, inherent value is equal in all entities with internal value, which also means that both moral patients and moral agents have internal equal value. Aldo Leopold believes that inherent value is related to the concept of social community, and entities have their moral status only in the interaction of relational bodies (Moline, [<reflink idref="bib33" id="ref44">33</reflink>]).</p> <p>Moral consideration involves the relations between various entities that depend on and adapt to each other in human society. The social relations among human beings, animals, machines, or AI, and the social structure that forms those relations, all constitute an important basis for human beings to give moral consideration to non-human entities. The popularity of intelligent technology has further blurred the boundary between human and AI. Ubiquitous intelligent devices have coexisted with human society, and their status depends on the relations with other entities. However, using AI to replace or enhance the functions of human body (wearable, implantable, and connectable) will not reduce the right of moral identification of human beings. The only problem may be the boundary. As AI and robots become more and more advanced and integrated into human society, human beings will be forced to reevaluate moral agents or moral patients. If AI shows the ability to reach or surpass human beings, the question will shift from "why should AI be given responsibilities and rights?" to "Should AI continue to be denied any moral consideration?." This is strikingly similar to the history of human struggle for equality.</p> <p>In short, at least since Aristotle's time, human beings have tried to explain and deal with differences by recognizing common point under external diversity. For example, traditional ethics put forward the common humanity, which is the basis of perception differences of race, gender, or class. Similarly, the centered biological ethics assumes that life itself has a common value, which contains all biodiversity. Although the purpose and use of machines are built by humans, their structure, objectives, and value depend on human purposes; however, Warren ([<reflink idref="bib50" id="ref45">50</reflink>]) still believes that "for complex electronic devices developed in the name of AI, the differences between mechanisms and organisms are likely to be unsustainable," future moral theories need to treat these excluded entities as equal legal moral patients. Himma ([<reflink idref="bib22" id="ref46">22</reflink>]) proposed that every existing entity, whether sentimental or merciless, living or inanimate, natural or artificial, has the minimum moral value by being. This gradual expansion strategy towards comprehensive understanding aims to put everything else down to something in common by gradually lowering the level of abstraction, thus transcending the dualistic concept defined by agent and patient. It also means that morality may not exist in moral agents or moral patients, but in their relations or the whole ecology.</p> <hd id="AN0161191385-8">The Educational Application of Moral AI</hd> <p>Moral consideration in the development of intelligent education system is becoming one of the important fields of AI research. If AI acts autonomously and the reason is based on moral reasoning, then the intelligent agent responsible for moral part is considered as a moral agent. Before robots become teachers and enter the classroom, it is necessary to endow them with morality. For example, what basic norms of teacher-student communication should AI teachers follow? How to play the expected role of teacher in the classroom? How do AI teachers respond to key aspects of educational interaction, and what decisions or when to act to affect learning outcomes? And so on.</p> <p>It is a complex and systematic project to realize moral AI in education. To be a "trustworthy" and "responsible" partner of teachers and students, educational AI must be widely consistent with teachers and students in terms of moral theoretical basis and expected value. Here, traditional ethics can play an important role. For example, deontology refers to individual behavior following certain moral principles or acting according to certain legitimacy. It is a normative moral stance, judges the morality of behavior by rules. One of the greatest advantages of deontology is that it can provide rational response in AI and is very transparent. Therefore, deontology is the fastest and simplest in a system small enough to be fully predictable, because all situations can be predicted. It can place the expected moral behavior in the traditional ethical category (prohibition, permission, and coercion) through the consistency test of behavior and rules (Kim et al., [<reflink idref="bib26" id="ref47">26</reflink>]). So, there will be no violation of the rules. For example, we use intelligent dialog agent to guide students to master the skills of recruitment interview. The deontological design includes two key parts: first, define detailed interview structure specifications and design different interview scenarios. These scenarios are labeled by detailed interview questions and ideal response behavior. Second, determine a detailed list to define and rank learners' observable behaviors (including different combinations and performances) according to importance.</p> <p>However, the above design cannot foresee all situations. Such as, the intelligent agent can neither foresee all learners' words and deeds, nor accurately define all observable words and deeds. While using an iterative algorithm such as the expectation–maximization algorithm can perform maximum likelihood estimation in the problem of incomplete data. It works by using existing values to calculate expected values and filling in missing values in the data (Zheng, [<reflink idref="bib53" id="ref48">53</reflink>]). So, if the behavior goal is clear enough and the intelligent agent can always fetch the current state, the utilitarian-based calculation model can reason and calculate all the expected results of the behavior to maximize the expectation or achieve a specific goal.</p> <p>Obviously, these premises cannot be fully met in the educational scene, because not only will there be a lot of incomplete, noisy, or erroneous information, but also all possible consequences of the behavior cannot be calculated (decisions must be made immediately, it is impossible to evaluate all possibilities). As Jeremy Bentham said, the process of hedonic calculus should not be expected to be strictly followed in every moral judgment (Schneider, [<reflink idref="bib39" id="ref49">39</reflink>]). For example, in the classroom, AI teachers should immediately decide what acceptable or unacceptable behavior is, AI must make the current optimal decision to determine the degree of evaluation or reasoning.</p> <p>Virtue ethics does not agree that moral behavior is the result of rules or consequence calculation. It considers broader cognitive functions, such as attention, emotion, and learning. Virtue ethics holds that good behavior comes from good character, and the development of character is a bottom-up experience learning process. So, only after learning and training can AI obtain stable strategies. To some extent, learning moral behavior is often accompanied by moral failure, so it is also necessary to evaluate the potential risk of allowing moral learning curve in autonomous machines. Simulation for some moral situations is still the best choice to train machines without actual harm.</p> <p>Some studies believe that the design method of moral learning is limited, because it can only reach the moral decision-making of human-level at most. Therefore, the "morality" of educational AI will never surpass teachers' morality (Chaudhuri & Vardi, [<reflink idref="bib13" id="ref50">13</reflink>]). But in fact, AI integrates a variety of moral theories, and through machine learning and multi-agent aggregation, special moral mistakes made by individuals can be marked in integration, it may produce a better moral system than any individual. Moreover, machine learning can discover the biases and limitations of human morality from the training set, and can also identify the general principles of moral decision-making that human beings were not aware of before. These principles can be used to improve human moral intuition. At present, moral AI is still in its infancy and is mostly used in specific fields. But in the future, it is foreseeable and hopeful to transfer moral reasoning from programmer to intelligent system itself, get rid of the current situation that it is too easy to embed moral reasoning into computing engine, and create a universal automatic moral decision-making intelligence with human level.</p> <p>The application of AI in education has just begun. The effective combination of machine intelligence and human intelligence is a significant educational innovation activity closely related to the development of science and technology. Educational robot involves multi-subjects (Wang et al., [<reflink idref="bib49" id="ref51">49</reflink>]); this integration promotes the development of humanization and individualized teaching, which will bring profound influence on education.</p> <hd id="AN0161191385-9">Discussion and Future Research</hd> <p>At present, human society is facing a major challenge about the moral consideration of AI. The rapid development of autonomous technology has had a great impact on human understanding of moral agent. As the decision-making and behavior of AI begin to change from external setting to internal development, intelligent agents gradually have different degrees of adaptive ability, behavior ability, and decision-making ability, and their autonomous learning ability is continuously enhanced. The role of AI in education is becoming more and more prominent, and the technologies and equipment driven by AI begin to enter the classroom. Compared with traditional teaching, this new form of collaborative teaching with AI participation redefines the mutual roles of students, teachers, and other actors.</p> <p>This has created a new situation in moral issues, and the traditional way of responsibility attribution conflicts with the moral framework. Any definition of moral agent or moral patient itself is a decisive moral consequence that predetermined who has or does not have moral status. Traditional ethics places human beings at the center, while animal ethics focuses on animals. During the process of development, environmental ethics goes further and considers all life, while existentialism expands the scope of moral consideration by combining with anything that already, actual, or possible existence, and provides the most universal and comprehensive definition of inclusive ethics in this way.</p> <p>Although these innovations are different, they all adopt a similar strategy of identifying common features underperformance diversity. According to Habermas ([<reflink idref="bib18" id="ref52">18</reflink>]), tolerance, whether it is another person, animal, environment, machine, or others, always and inevitably encounters the same methodological problem, namely, reducing differences. Therefore, if moral considerations are extended to other traditionally marginalized entities, more inclusive moral standards need to be gradually defined.</p> <p>The evolving moral theory redefines (deconstructs) the center of moral considerations, effectively reduces differences, becomes more inclusive, and includes more possible participants. But maybe we need to jump out of the binary framework and solve the problem by rewriting rules. For example, Hanson ([<reflink idref="bib21" id="ref53">21</reflink>]) introduces a new kind of moral agent, i.e., extended agencies. The responsibility does not lie in the predetermined moral individuals, but the network of relationships between individuals and others including machines. Cybernetic ecology believed that "in the ecosystem, all elements have a certain inherent value, but due to the interdependence within the system, each element also has a certain tool value for other parts of the system. Each part of the ecology has a certain level of autonomy. But in the system environment, each part only plays a certain role in controlling the entire ecology" (Channell, [<reflink idref="bib12" id="ref54">12</reflink>]). Each element becomes the product of its position in the whole and is endowed with appropriate rights and responsibilities.</p> <p>Human society usually judges whether the moral agent is in line with the moral norms according choosing the behavior and its consequences. As expected, the behavior of the moral agent will produce good results in morality. However, AI's behavior is still somewhat uncertain; sometimes the behavior will not necessarily achieve the desired results, or go wrong, which is particularly important for AI teachers. Only when educational AI obtains moral attributes, its educational goals, decisions, and words or actions taken to achieve the goals will be consistent with human expected morality and values.</p> <p>AI's responsibilities and rights will be established after solving the major philosophical problems, and the application premise of AI in education will be met. At present, AI technology has been applied in various educational scenarios to varying degrees (Xu et al., [<reflink idref="bib52" id="ref55">52</reflink>]). In the future, there will definitely be a "double teacher classroom" composed of AI teachers and human teachers. This will completely change the way of teaching and learning, and lead to a revolution in education. Therefore, the deep integration of AI and education will become the development trend of education in the future.</p> <hd id="AN0161191385-10">Declarations</hd> <p></p> <hd id="AN0161191385-11">Conflict of Interest</hd> <p>The authors declare that they have no conflict of interest.</p> <hd id="AN0161191385-12">Publisher's Note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0161191385-13"> <title> References </title> <blist> <bibl id="bib1" idref="ref28" type="bt">1</bibl> <bibtext> Arthur WB. The nature of technology: What it is and how it evolves. 2009; Simon and Schuster</bibtext> </blist> <blist> <bibl id="bib2" idref="ref16" type="bt">2</bibl> <bibtext> Baber, C. (2003). Cognition and tool use: Forms of engagement in human and animal use of tools. Routledge-Taylor & Francis Group.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref42" type="bt">3</bibl> <bibtext> Baujard A. A return to Bentham's felicific calculus: From moral welfarism to technical non-welfarism. The European Journal of the History of Economic Thought. 2009; 16; 3: 431-453. 10.1080/09672560903101294</bibtext> </blist> <blist> <bibl id="bib4" idref="ref1" type="bt">4</bibl> <bibtext> Beauchamp TL, Frey RG. The Oxford handbook of animal ethics. 2011; Oxford University Press. 10.1093/oxfordhb/9780195371963.001.0001</bibtext> </blist> <blist> <bibl id="bib5" idref="ref34" type="bt">5</bibl> <bibtext> Bostrom N, Yudkowsky EKeith F, William MR. The ethics of artificial intelligence. The Cambridge handbook of artificial intelligence. 2014; Cambridge University Press</bibtext> </blist> <blist> <bibl id="bib6" idref="ref19" type="bt">6</bibl> <bibtext> Bryson JJWilks Y. Robots should be slaves. Close engagements with artificial companions: Key social, psychological, ethical and design issues. 2010; John Benjamins Publishing Company</bibtext> </blist> <blist> <bibl id="bib7" idref="ref15" type="bt">7</bibl> <bibtext> Burri, T, & Trusilo, D. (2021). Ethical artificial intelligence: An approach to evaluating disembodied autonomous systems. In Rain, L, & Ann, V. (Eds), Autonomous cyber capabilities in international law, forthcoming. Available on SSRN: https://ssrn.com/abstract=3816687.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref17" type="bt">8</bibl> <bibtext> Butler, S. (1912). The note-books of Samuel Butler. (n.d.). Retrieved January 20,2021, from <ulink href="http://www.public-library.uk/pdfs/7/933.pdf">http://www.public-library.uk/pdfs/7/933.pdf</ulink>.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref38" type="bt">9</bibl> <bibtext> Calarco M. Zoographies: The question of the animal from Heidegger to Derrida. 2008; Columbia University Press</bibtext> </blist> <blist> <bibtext> Calverley DJ. Android science and animal rights, does an analogy exist?. Connection Science. 2006; 18; 4: 403-417. 10.1080/09540090600879711</bibtext> </blist> <blist> <bibtext> Cavalier RJ. Impact of the internet on our moral lives. 2005; State University of New York Press</bibtext> </blist> <blist> <bibtext> Channell DF. The vital machine: A study of technology and organic life. 1991; Oxford University Press</bibtext> </blist> <blist> <bibtext> Chaudhuri, S, & Vardi, M. (2014). Reasoning about machine ethics. Principles of programming languages (POPL)-off the beaten track (OBT). Retrieved January 20,2021, from https://popl-obt-2014.cs.brown.edu/papers/ethics.pdf.</bibtext> </blist> <blist> <bibtext> Descartes R. Descartes: Selected philosophical writings. 1988; Cambridge University Press. 10.1017/CBO9780511805059</bibtext> </blist> <blist> <bibtext> Floridi L. The Cambridge handbook of information and computer ethics. 2010; Cambridge University Press. 10.1017/CBO9780511845239</bibtext> </blist> <blist> <bibtext> Floridi L. The ethics of information. 2013; Oxford University Press. 10.1093/acprof:oso/9780199641321.001.0001</bibtext> </blist> <blist> <bibtext> Floridi L, Sanders JW. On the morality of artificial agents. Minds and machines. 2004; 14; 3: 349-379. 10.1023/B:MIND.0000035461.63578.9d</bibtext> </blist> <blist> <bibtext> Habermas J. The inclusion of the other: Studies in political theory. 1998; MIT Press</bibtext> </blist> <blist> <bibtext> Haenlein M, Kaplan A. A brief history of artificial intelligence: On the past, present, and future of artificial intelligence. California Management Review. 2019; 61; 4: 5-14. 10.1177/0008125619864925</bibtext> </blist> <blist> <bibtext> Hajdin M. The boundaries of moral discourse. 1994; Loyola University Press</bibtext> </blist> <blist> <bibtext> Hanson FA. Beyond the skin bag: On the moral responsibility of extended agencies. Ethics and Information Technology. 2009; 11; 1: 91-99. 10.1007/s10676-009-9184-z</bibtext> </blist> <blist> <bibtext> Himma KE. There's something about Mary: The moral value of things qua information objects. Ethics and Information Technology. 2004; 6; 03: 145-159. 10.1007/s10676-004-3804-4</bibtext> </blist> <blist> <bibtext> Howard D, Muntean IPowers TM. Artificial moral cognition: Moral functionalism and autonomous moral agency. Philosophy and computing. 2017; Springer</bibtext> </blist> <blist> <bibtext> Ikäheimo, H, & Arto, L. (2007). Dimensions of personhood: Editors' introduction. Journal of Consciousness Studies,14(5–6), 6–16. Retrieved January 20,2021,from <ulink href="http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.63.3611&rep=rep1&type=pdf">http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.63.3611&rep=rep1&type=pdf</ulink>.</bibtext> </blist> <blist> <bibtext> Kant I. Lectures on ethics. 1997; Cambridge University Press. 10.1017/CBO9781107049512</bibtext> </blist> <blist> <bibtext> Kim TW, Hooker J, Donaldson T. Taking principles seriously: A hybrid approach to value alignment in artificial intelligence. Journal of Artificial Intelligence Research. 2021; 70: 871-890. 10.1613/jair.1.12481</bibtext> </blist> <blist> <bibtext> Lindemann G. Eccentric positionality: On Kant, Plessner, and human dignity. An interview with JM Bernstein. Human Studies. 2019; 42; 1: 147-158. 10.1007/s10746-018-09486-z</bibtext> </blist> <blist> <bibtext> MacIntyre AC. Dependent rational animals: Why human beings need the virtues. 1999; Open Court Publishing</bibtext> </blist> <blist> <bibtext> Marx K. Capital: A critique of political economy. 1977; Vintage Books</bibtext> </blist> <blist> <bibtext> Misselhorn, C. (2018). Artificial morality, concepts, issues and challenges. Society, 55(2), 161–169. https://doi.org/10.1007/s12115-018-0229-y.</bibtext> </blist> <blist> <bibtext> McLuhan M. Understanding media: The extensions of man. 1995; MIT Press</bibtext> </blist> <blist> <bibtext> Merleau-Ponty. M. (2006). The structure of behavior (8.print.). Duquesne Univ. Press.</bibtext> </blist> <blist> <bibtext> Moline JN. Aldo Leopold and the moral community. Environmental Ethics. 1986; 8; 02: 99-120. 10.5840/enviroethics19868222</bibtext> </blist> <blist> <bibtext> Moor JH. The nature, importance, and difficulty of machine ethics. IEEE Intelligent Systems. 2006; 21; 4: 18-21. 10.1109/MIS.2006.80</bibtext> </blist> <blist> <bibtext> Naas M. Plato and the invention of life. 2018; Fordham University Press. 10.2307/j.ctt2204pw8</bibtext> </blist> <blist> <bibtext> O'Hagan E. Animals, agency, and obligation in Kantian ethics. Social Theory and Practice. 2009; 35; 4: 531-532. 10.5840/soctheorpract200935431</bibtext> </blist> <blist> <bibtext> Rowlands M. Animal rights: Moral theory and practice. 2009; Palgrave Macmillan</bibtext> </blist> <blist> <bibtext> Salehi, A. A, & Talebzade, H. (2020). Hegel's metaphysical formulation of labor in phenomenology of spirit and its critique in Marx by introducing praxis as the new foundation. Journal of Philosophical Investigations, 14(30), 142–164. https://doi.org/10.22034/JPIUT.2020.35824.2411.</bibtext> </blist> <blist> <bibtext> Schneider RC. Developing moral sport policies through act utilitarianism based on Bentham's hedonic calculus. Sport Management International Journal. 2010; 6; 2: 51-62. 10.4127/ch.2010.0051</bibtext> </blist> <blist> <bibtext> Scott PA. Morally autonomous practice?. Advances in Nursing Science. 1998; 21; 2: 69-79. 10.1097/00012272-199812000-00008</bibtext> </blist> <blist> <bibtext> Searle JR, Dennett DC, Chalmers DJ. The mystery of consciousness. 1997; The New York Review of Books</bibtext> </blist> <blist> <bibtext> Shapiro P. Moral agency in other animals. Theoretical Medicine and Bioethics. 2006; 27; 4: 357-358. 10.1007/s11017-006-9010-0</bibtext> </blist> <blist> <bibtext> Sharkey A. Should we welcome robot teachers?. Ethics and Information Technology. 2016; 18; 4: 283-297. 10.1007/s10676-016-9387-z</bibtext> </blist> <blist> <bibtext> Singer, P. (2007). All animals are equal. In Chadwick, R, & Schroeder, D. (Eds). Applied ethics: Critical concepts in philosophy. Routledge-Taylor & Francis Group.</bibtext> </blist> <blist> <bibtext> Steiner P, Stewart J. From autonomy to heteronomy (and back): The enaction of social life. Phenomenology and the Cognitive Sciences. 2009; 8; 4: 527-529. 10.1007/s11097-009-9139-1</bibtext> </blist> <blist> <bibtext> Torrance S. Ethics, consciousness and artificial agents. AI and Society. 2008; 22; 4: 495-521. 10.1007/s00146-007-0091-8</bibtext> </blist> <blist> <bibtext> Waal FD. Good natured: The origins of right and wrong in humans and other animals. 1996; Harvard University Press. 10.4159/9780674033177</bibtext> </blist> <blist> <bibtext> Wallach W, Allen C. Moral machines: Teaching robots right from wrong. 2009; Oxford University Press. 10.1093/acprof:oso/9780195374049.001.0001</bibtext> </blist> <blist> <bibtext> Wang, B, Liu, H, & An, P. et al. (2018). Artificial intelligence and education. In Jin, D. (Ed.) Reconstructing our orders. Springer. https://doi.org/10.1007/978-981-13-2209-9_5.</bibtext> </blist> <blist> <bibtext> Warren TP. Respect for nature: A theory of environmental ethics. 1986; Princeton University Press</bibtext> </blist> <blist> <bibtext> Winner L. Autonomous technology: Technics-out-of-control as a theme in political thought. 1977; MIT Press</bibtext> </blist> <blist> <bibtext> Xu, Z, Wei Y, & Zhang, J. (2021). AI applications in education. In Shi, S, Ye, L, &Zhang, Y. (Eds.) Artificial intelligence for communications and networks. AICON 2020. Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering, vol 356. Springer. https://doi.org/10.1007/978-3-030-69066-3_29.</bibtext> </blist> <blist> <bibtext> Zheng, Y. (2011). Machine learning with incomplete information. PhD dissertation. The University of Nebraska-Lincoln.</bibtext> </blist> </ref> <aug> <p>By Fuhai Sun and Ruixing Ye</p> <p>Reported by Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib38" firstref="ref2"></nolink> <nolink nlid="nl2" bibid="bib19" firstref="ref3"></nolink> <nolink nlid="nl3" bibid="bib43" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib28" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib44" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib24" firstref="ref7"></nolink> <nolink nlid="nl7" bibid="bib14" firstref="ref8"></nolink> <nolink nlid="nl8" bibid="bib36" firstref="ref9"></nolink> <nolink nlid="nl9" bibid="bib42" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib30" firstref="ref11"></nolink> <nolink nlid="nl11" bibid="bib47" firstref="ref12"></nolink> <nolink nlid="nl12" bibid="bib10" firstref="ref13"></nolink> <nolink nlid="nl13" bibid="bib32" firstref="ref14"></nolink> <nolink nlid="nl14" bibid="bib35" firstref="ref18"></nolink> <nolink nlid="nl15" bibid="bib31" firstref="ref20"></nolink> <nolink nlid="nl16" bibid="bib12" firstref="ref21"></nolink> <nolink nlid="nl17" bibid="bib41" firstref="ref22"></nolink> <nolink nlid="nl18" bibid="bib11" firstref="ref23"></nolink> <nolink nlid="nl19" bibid="bib40" firstref="ref24"></nolink> <nolink nlid="nl20" bibid="bib51" firstref="ref25"></nolink> <nolink nlid="nl21" bibid="bib45" firstref="ref26"></nolink> <nolink nlid="nl22" bibid="bib29" firstref="ref27"></nolink> <nolink nlid="nl23" bibid="bib16" firstref="ref29"></nolink> <nolink nlid="nl24" bibid="bib27" firstref="ref30"></nolink> <nolink nlid="nl25" bibid="bib17" firstref="ref31"></nolink> <nolink nlid="nl26" bibid="bib48" firstref="ref32"></nolink> <nolink nlid="nl27" bibid="bib46" firstref="ref33"></nolink> <nolink nlid="nl28" bibid="bib34" firstref="ref35"></nolink> <nolink nlid="nl29" bibid="bib23" firstref="ref36"></nolink> <nolink nlid="nl30" bibid="bib20" firstref="ref37"></nolink> <nolink nlid="nl31" bibid="bib37" firstref="ref39"></nolink> <nolink nlid="nl32" bibid="bib15" firstref="ref40"></nolink> <nolink nlid="nl33" bibid="bib25" firstref="ref43"></nolink> <nolink nlid="nl34" bibid="bib33" firstref="ref44"></nolink> <nolink nlid="nl35" bibid="bib50" firstref="ref45"></nolink> <nolink nlid="nl36" bibid="bib22" firstref="ref46"></nolink> <nolink nlid="nl37" bibid="bib26" firstref="ref47"></nolink> <nolink nlid="nl38" bibid="bib53" firstref="ref48"></nolink> <nolink nlid="nl39" bibid="bib39" firstref="ref49"></nolink> <nolink nlid="nl40" bibid="bib13" firstref="ref50"></nolink> <nolink nlid="nl41" bibid="bib49" firstref="ref51"></nolink> <nolink nlid="nl42" bibid="bib18" firstref="ref52"></nolink> <nolink nlid="nl43" bibid="bib21" firstref="ref53"></nolink> <nolink nlid="nl44" bibid="bib52" firstref="ref55"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1361283
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Moral Considerations of Artificial Intelligence
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Sun%2C+Fuhai%22">Sun, Fuhai</searchLink><br /><searchLink fieldCode="AR" term="%22Ye%2C+Ruixing%22">Ye, Ruixing</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Science+%26+Education%22"><i>Science & Education</i></searchLink>. Feb 2023 32(1):1-17.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 17
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2023
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Descriptive
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Moral+Issues%22">Moral Issues</searchLink><br /><searchLink fieldCode="DE" term="%22Ethics%22">Ethics</searchLink><br /><searchLink fieldCode="DE" term="%22Artificial+Intelligence%22">Artificial Intelligence</searchLink><br /><searchLink fieldCode="DE" term="%22Definitions%22">Definitions</searchLink><br /><searchLink fieldCode="DE" term="%22Inclusion%22">Inclusion</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Technology%22">Educational Technology</searchLink><br /><searchLink fieldCode="DE" term="%22Technology+Integration%22">Technology Integration</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Trends%22">Educational Trends</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1007/s11191-021-00282-3
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0926-7220<br />1573-1901
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: One of the ultimate problems of moral philosophy is to determine who or what is worth moral consideration or not. "Morality" is a relative concept, which changes significantly with the environment and time. This means that morality is incredibly inclusive. The emergence of AI technology has a significant impact on the understanding and distribution of "subject," which has produced a new situation in moral issues. When considering the morality of AI, moral problems must also involve moral agents and moral patients. A more inclusive moral definition is necessary for extending the scope of moral consideration to other traditionally marginalized entities. The evolving ethics redefines the center of moral consideration, effectively reduces the differences, becomes more inclusive, and includes more potential participants. But we may still need to jump out of this binary framework and solve the problem by rewriting rules. It is a huge, complex systematic project to realize moral AI in education. To be a "trustworthy" and "responsible" companion of teachers and students, educational AI must have extensive consistency with teachers and students in terms of the moral theoretical basis and expected value. The deep integration of AI and education is likely to become the development trend of education in the future.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2023
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1361283
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1361283
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s11191-021-00282-3
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 17
        StartPage: 1
    Subjects:
      – SubjectFull: Moral Issues
        Type: general
      – SubjectFull: Ethics
        Type: general
      – SubjectFull: Artificial Intelligence
        Type: general
      – SubjectFull: Definitions
        Type: general
      – SubjectFull: Inclusion
        Type: general
      – SubjectFull: Educational Technology
        Type: general
      – SubjectFull: Technology Integration
        Type: general
      – SubjectFull: Educational Trends
        Type: general
    Titles:
      – TitleFull: Moral Considerations of Artificial Intelligence
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Sun, Fuhai
      – PersonEntity:
          Name:
            NameFull: Ye, Ruixing
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Type: published
              Y: 2023
          Identifiers:
            – Type: issn-print
              Value: 0926-7220
            – Type: issn-electronic
              Value: 1573-1901
          Numbering:
            – Type: volume
              Value: 32
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Science & Education
              Type: main
ResultId 1