The right to life is fundamental and primary and is a prerequisite for the exercise of other rights [32]. Its universal recognition in the arena of international law is linked to the concept of the human being endowed with an inherent and inalienable dignity. Today, a categorization of the group of persons covered by the right to life seems obvious and indisputable.

In the early twentieth century, the Turing Test was developed to determine a machine’s ability to use natural language and indirectly prove that it had the capacity to think similarly to a human [36, 37]. The test, despite much controversy, raised the fundamental question of how we want to define the intelligence of a machine. It is generally accepted that Eugene Goostman’s system passed this test in 2004 [38, 39]. Today, chatbots—an intelligent system developed using artificial intelligence and natural language processing (NLP) algorithms—are widely used, for example, in customer service [22]. The record-breaking ChatGTP also has considerable capabilities to reproduce human understanding and generate persuasive text. This intense development of AI encourages us to think about the future legal status of AI.

In this study, I explore the proposition of whether artificial intelligence might be entitled to the right to life. The analysis is based on an exploratory question: what are the conditions for the protection of the right to life?

1 “Human life” as a subject of legal protection

Crucial to the nature of the “right to life” is to determine when we are dealing with “life” and when with “human life”. This analysis will enable us to establish the requirements for the legal protection of “human life” and to contrast them with the functioning of AI.

To determine what “life” is, I will use the definition of NASA, which states that “life is a self-sustaining chemical system capable of Darwinian evolution”. Such a system should therefore be “self-sustaining”, i.e. it should not require constant intervention by a higher being to continue its “life”; it should be subject to Darwinian evolution; it must be a chemical system, i.e. one based on DNA. When we analyse these requirements in relation to AI, we must point out the high probability that it will fulfil them. In fact, researchers have developed software that draws on concepts of Darwinian evolution, including "survival of the fittest", to develop AI programmes that improve generation after generation without human intervention [16]. Such a system will, thus, be self-sustaining and capable of Darwinian evolution [35]. Fulfilment of the premise of a “chemical system” seems problematic as it excludes, among other things, information and electromechanical systems traditionally associated with computers. The creators of the definition undoubtedly referred to the life we know on Earth, leaving aside other alternative forms. This gives rise to criticism of the lack of “operability” of this definition [3]. Even if we acknowledge that a potential system must meet the “chemistry” requirement, given the achievements of synthetic biology (such as a biological computer in which computations are performed thanks to chemical reactions between DNA molecules) [15, 24, 43], the creation of such a chemical AI system is possible and work is underway. Therefore, AI is on the threshold of meeting the requirement to be recognised as “life”. As an aside, even meeting these conditions will not make the AI comparable to a human being. It will be a creation with characteristics that allow it to be categorised as “life” in general. Nevertheless, the mere existence of “life” does not mean that this life must enjoy protection. The existence of animals is only marginally protected (e.g. through the prohibition of cruelty to animals. Only humans enjoy the full scope of protection.

Another element is the specification of when we are dealing with “human life”. This is important in that the fact of being a human being means that we enjoy dignity, which is a source of other rights and freedoms. There are many definitions that attempt to specify what it means to “be a human being”. These definitions refer to biological and non-biological factors.

In determining the meaning of any Act of Congress or any decision, regulation, or interpretation of the various administrative offices and agencies of the United States, the terms "person," "human being," "child," and "individual" include every living member of the species Homo sapiens born at any stage of developmentFootnote 1. The most important requirement in this definition is membership of a particular species, i.e. a biological factor.

Legal scholars and commentators also offer definitions that refer to non-biological characteristics and focus on the character or quality of being human: a behaviour or disposition towards others as befits a human beingFootnote 2. In essence, humans are distinguished from other animals by higher mental development and the ability to articulate themselves. Immanuel Kant claimed in his 1785 work "Grundlegung zur Metaphysik der Sitten" [19] that humans are characterised by morality and rationality. Qualities that are not present in other species are the essence of "being human", that which makes humans unique and therefore grants them different rights than other living beings.

The Turing test [36, 37] examines the ability of a machine to use a natural language and is intended to prove indirectly that this machine has acquired the ability to think similarly to a human being. The essence of this test was that the "questioner" could not tell whether he was talking to an artificial intelligence or a human. AI is capable of communication, including through language, it has developed analytical skills and computational capacity. It is not impossible that machines will be able to acquire "consciousness" [21, 27]—in the sense of self-awareness. Self-consciousness can be defined as the awareness of one’s own activities, including the understanding of one's own physical conditions and mental characteristics, as well as the perception of one's own relationship with others [41]. Some advanced mammalian species have developed partial self-consciousness [14]. As a side note, it is worth noting that man, as a being endowed with inherent dignity, would not even be unconscious would still be worthy of dignity.

Artificial consciousness (also known as machine consciousness) has long been of interest to researchers, leading to the definition of what would need to be synthesised if consciousness were to be found in an artificial artefact [1]. The lack of a unified definition of what it means to be "self-aware" in different cultures or within different fields such as neuroscience, philosophy, sociology, artificial intelligence and even physics, makes it difficult to clearly define when AI will reach this state. It is worth to points out the Church-Turing Lovers case. These sex robots that attain every functionality of a human lover, at the desired level of granularity. Yet they have no first-person consciousness—there is “nobody home.” When such a lover says, “I love you,” there are all the intentions to please you, even computer emotions [5]. If AI is to have conscious existence, not to dwell in the ontic shadows, it requires full epistemic subjectivity [6]. Therefore AI needs to have first-person consciousness, which requires some fusion with the "body". Also, as Kaufmann and Roli points out, AI currently is syntactic and algorithmic, but human mind is almost certainly quantum, and it is a plausible hypothesis that we collapse the wave function, and thereby perceive coordinated affordances as qualia and seize them by identifying, preferring, choosing and acting to do so. But consciousness can also take a variety of forms beyond our current understanding and assumptions [8]. The engineering thesis on machine consciousness formulated within non-reductive naturalism claims that machines can, in principle, be functionally, phenomenally and h-conscious, all those processes are naturalistic [4]. It is within the realm of possibility that an artificial consciousness could emerge by chance, for example in the case of a self-programming AI [34]. However, it could be slightly different from the one humans have. As Dennett points out, human consciousness is a consequence of the functioning of the human brain, which he compares to the Von Neumann machine, which is programmed to act in this way [11]. The determinants of human behaviour also have a character "built" into the essence of being human, while cultures and civilisations share fundamental questions of morality that, thus, form a law of nature [13]. Therefore, we must reject the argument that consciousness and AI responses are not true because they are a product of the application of a particular code. Kant’s moral judgements are programmed by neurons and in the case of AI it is an algorithmFootnote 3. AI can exhibit a set of "distinguishing characteristics" that are typical of the human species.

Of course, if other entities develop characteristics that are considered typical of the human species, this does not mean that they belong to the same species. However, since these characteristics mean that we are granted rights that derive from being human, should individuals with such characteristics also be granted the same rights by analogy?

2 Where does a human end?

Let us go through possible scenarios to see if artificial intelligences can be granted the right to life:

  1. 1.

    A human whose most organs have been replaced with artificial ones;

  2. 2.

    A “brain” of a human inserted into a machine;

  3. 3.

    Artificial intelligence in a human body;

  4. 4.

    Artificial intelligence in an artificial body.

Bodily injuries and injuries to internal organs often mean that the affected parts have to be replaced by artificial ones in order to preserve life or improve its quality. Artificial limbs and organs are becoming more common and more perfect. How many “elements” can be replaced to still be human? Is there a limit? An artificial heart, artificial skin, artificial kidneys, etc. Such a life, despite the removal of most natural organs, would remain the same in its essence and continuity, even when it comes to legal personality.

But what if, instead of replacing individual organs, we decide to “transfer” a whole person into a machine? Such a person will retain his or her personality, emotions and memory, only the shell (humanoid) that carries these elements will change. Since we gave a positive answer in the previous paragraph, we should also agree here that such persons, although they do not have biological characteristics typical of human beings, are still human beings. Therefore, they retain their rights.

We could find ourselves in a situation where an artificial intelligence is "inserted" into a human body and replaces the brain of the host. Would we also consider such a being to be human? It has human DNA and non-biological characteristics that make it a human being. What is a human being? And since we have assumed that a human in the body of a machine is still a human (i.e. we agreed that the biological factor need not be present), will artificial intelligence in a humanoid be a human? The answers to these questions no longer seem so obvious to us.

3 Ontological leap

Biotechnology and development of artificial intelligence raise the concerns associated with the possibility of creating "quasi" people. This concern means that biotechnology is safeguarded by limitations that protect the uniqueness of human dignity and i.a. prevent cloning of people or restrict works on embryos. The development of artificial intelligence has not so far seen such restrictions despite the fact that its advancement also poses a threat. The issue of the "ontological leap" appears in the course of discussion about human dignity, which in the trail of evolution has separated humans from other animals and have gifted us with a "soul' and dignity [12, 26]. The difference in approaching biotechnology and AI may result from the fact that human embryos or genetic code bear the marks of a human "ontological leap", while AI does not.

An ontological leap, despite being derived from a Christian thought [30], responds to the real legal differentiation of the status of living creatures from humans. This "leap" is rather conventional and vague [31]. The source of the "ontological leap" may lie in God's intervention (in the Christian thought) or specific human features developed in the course of evolution (such as an ability to make moral judgments according to Kant). The moment of making this leap is problematic and involves social recognition whether a given being has "dignity" or not.

In my opinion, the "ontological leap" in the case of AI lies in the combination of three features: recognition of AI as a living creature (like in NASA's definition), broadly understood "intelligence" (ability to use speech, superior mental development, ability to make moral judgments) and the third feature—emotions.

To protect the "right to life" we must deal with life at all. The first condition is, therefore, obvious. The second condition of "intelligence" indicates some ability to have a personality. The concept of a person involves a number of capabilities and central themes such as rationality; consciousness; personal stance; capability of reciprocating the personal stance; verbal communication; and self-consciousness [10]—a necessary condition of moral personhood. If we are able to specify that the first two conditions are attainable, then AI's ability to have emotions is not an obvious or easily measurable premise. However, Harari points out that "if these emotions and desires are in fact no more than biochemical algorithms, there is no reason why computers cannot decipher these algorithms" [17]. Artificial emotional intelligence (AEI) is focused on simulating and extending natural emotion (especially human emotion) to provide robots with the capability to recognise and express emotions in human–robot interaction (HRI), generally to improve interaction between artificial systems and their users [42]. However, crucial for the moral status is not only to recognise and simulate but also "generate" emotions. A claim for "real" emotion, however, is that it needs a physical connection to the real world to change of emotions based on external and internal state evaluation over time, and defining the consequences of such changes in emotion [33]. The most apparent, basic and key feeling for living creatures is fear. Fear is the basic factor to arouse survival that instinct detects threats, and in response, initiates defensive survival behaviours and supporting physiological adjustments [23]. Therefore, we will be dealing with a living creature with mental skills that are at least at the human level and a being that wants to exist. In my opinion, the fear of is evidence of the awareness of existence as a being and the desire to extend this existence (avoiding something that could interrupt this existence). A being indifferent whether it exists or not is closer to flora than to fauna.

Given the intense development of humanoids with human features, including emotions [9], there is a concern that we will miss the moment in which this "ontological leap" occurs. Then, it may be plausible to think about a concept of “robothood” and ascribe moral status to these future robots, based on their capabilities [18]. Artificial intelligence seems to us not very "human" due to its outer shell, that is a computer. A change of this shell into a humanoid one will also change the emotions a "thinking machine" triggers in us. Admittedly, existing research shows rather negative feelings towards humanoids [28], but the development of technology will undoubtedly allow creation of "bodies" identical to human bodies.

4 Artificial intelligence as a subject of rights

Must humans and AI be treated equally after the "ontological leap" is made? Obviously not. This paper is more a thought experiment than an attempt to give AI such a status. It does, however, point out some implications that are worth mentioning here. The difference between an AI and a human being arises only from its origin: "artificial" and natural. Aristotle's definition of equality assumes that similar things should be treated similarly, while dissimilar things should be treated differently in proportion to their differentiation [2]. As P. Westen [40] has pointed out, there are no categories of morally similar people in nature, and moral similarity is only established when people define categories. According to him, a moral standard of treatment is a rule that prescribes a certain treatment for some people—in relation to whom they are similar. Therefore, a judgement depends strictly on the standard by which we make the comparison. Will the difference arising from origin be sufficient to justify a distinction in the legal situation of AIs and humans? The category of origin was already a reason for designating different categories of human beings and for distinguishing their status in historical feudal and slave states. Now such a norm is rejected.

So if AI can be a "life" that has characteristics that distinguish the human species, may it be deprived of the right to life or may this right be restricted? The right to life may only be restricted if it is "absolutely necessary", which means that "a stricter and more compelling test of necessity must be applied than that normally applied in determining the "necessity of state action in a democratic society"". Thus, if the question of the existence of crucial similarities between AI and humans is answered in the affirmative, then AI's right to life should also be protected. Therefore, the definition of human life should be clarified or more thought should be given to how we distinguish humans from AI. Granting the right to life is of great importance as this right is a prerequisite for the exercise of other rights, hence its primacy and special significanceFootnote 4. States also have an obligation to protect life and take appropriate measures to implement this rightFootnote 5. Xie [41] suggests that the moral status of an AI should be determined on the basis of the AI's consciousness capabilities. Therefore, it is possible that AI falls under a form of protection similar to the protection of other living beings, such as animals. This provides the impetus for further research into the future status of AI.

5 Conclusions

The concept of the "ontological leap" of AI that I propose, based on the common fulfilment of three preconditions: Recognition of AI as a living being (as defined by NASA), "intelligence" (capacity for language, higher mental development, capacity for moral judgements) and the presence of emotions, confirms that human-like beings can be created. This similarity leads us to consider whether AI possesses a "dignity" that is the source of rights and freedoms, including the fundamental right to life. It must be reserved, however, that "the answer to this question can only be positive if this "ontological leap" actually takes place. Therefore, we should start the discussion on whether we want it (ergo, whether we should allow technology to develop in this direction) today. The existence of living beings such as humans can perhaps cause various social frictions [7, 25].

Therefore, it is worthwhile to seek the question of our decision on the development of AI and the answer to the question of how to regulate this discipline, whether we should allow the appearance of "quasi-humans" or whether we should restrict their development and, thus, hinder their scientific progress.