Use of Artificial Intelligence in the Activities of Religious Associations and Control Over Them

Introduction

At the “AI Journey” conference on November 24, 2022, the Russian President V. V. Putin noted the need to provide national projects and state program with specific measures aimed at the introduction of artificial intelligence technologies into industries. The results of using the artificial intelligence in various sectors of the economy and social sphere should be registered using monitoring and a special tool  – an index of maturity of industries and regions (“the index of intellectual maturity”), aimed at collecting information and assessing the performance of each Federation subject, ministry, or department on the introduction of artificial intelligence1.

The application of artificial intelligence in all spheres of human life will also affect the religious sphere, religious organizations and public authorities vested with regulatory and control (supervisory) powers in the sphere of religious organizations. Therefore, it is necessary to assess in advance the risks and problems that may be faced by religious organizations, society and the state in connection with the introduction of artificial intelligence in the spiritual sphere².

1. Practice and principles of using digital technologies in the activities of religious associations

In the field of jurisprudence, there are few publications that deal with the legal aspect of the artificial intelligence application in the activities of religious associations or digitalization of religious activities (Nathan & Walker, 2014; Akhmedov et al., 2021). Religion does not stand still but develops, adapting to changes in society, transformation of relations between the sexes (genders), climate change, etc. Some religions are “dying”, while others are emerging and flourishing. Traditional religions are being transformed by widespread digitalization in order to maintain and increase the number of their followers and to keep up with the times. Digitalization is changing traditional communication patterns (Smirnov, 2019), which are of particular importance in the religious sphere.

Not only secular authorities are concerned about formulating ethical and legal principles for the development, implementation and use of artificial intelligence, but also representatives of various religions. Information technologies, including artificial intelligence, are changing the attitude to religion, the way it is practiced and carried to the masses. Religions and churches that are not represented on the Internet are becoming disconnected from their parishioners.

Vatican has taken an active stance in elaborating ethical rules for the creation and application of artificial intelligence (Abramov, 2020). It was, actually, one of the first to introduce a “virtual” presence, noting the importance of the Internet for the Church³. For example, the function of sending e-mails to the Pope was first launched on the Vatican’s website in 2005, and earlier, in 1995, the Christmas message of John Paul II was disseminated via the Internet. In 2011, Pope Benedict XVI created a Twitter⁴ account.

In February 2019, Pope Francis, in a private meeting with Microsoft CEO Brad Smith, noted that artificial intelligence must serve the common good of humanity and must not allow for digital inequality. Following this meeting, the Pontifical Academy for Life and Microsoft agreed to establish an international award in the field of artificial intelligence⁵. In 2019 and 2020, conferences were held on “Roboethics. Humans, Machines and Health” and “The Good Algorithm? Artificial Intelligence: Ethics, Law, Health”, respectively (von Braun et al., 2021). Later, the Academy presented the “Rome Call for AI Ethics” (February, 2020):

• transparency: AI systems’ behavior and decision making must be explainable;
• inclusion: artificial intelligence must be useful for all human beings and offer all individuals the best possible conditions for self-expression and development;
• responsibility: artificial intelligence developers must act with responsibility;
• impartiality: artificial intelligence must have no bias but safeguard human dignity;
• reliability: artificial intelligence must be able to work reliably;
• security and privacy: artificial intelligence must work securely and respect the privacy of users⁶.

Somewhat earlier, in April 2019, the USA Soutern Baptist Convention formulated “An Evangelical Statement of Principles for artificial intelligence”⁷:

• artificial intelligence is a technology that should not be assigned a level of human identity, human values, dignity and moral freedom;
• artificial intelligence is not worthy of human hope, worship or love. Artificial intelligence should not devalue or degrade the dignity and worth of human beings;
• responsibility for the actions of artificial intelligence should rest solely with humans;
• artificial intelligence should not be used as a means to improve, change or perfect humans;
• artificial intelligence should be unbiased;
• the use of artificial intelligence for sexual objectives should not replace sexual relations between a man and a woman;
• humanity should not use artificial intelligence to move towards a life of pure leisure;
• it is inadmissible to use data inappropriately, with the objective of harming, humiliating, distorting reality, strengthening some people and humiliating others;
• the use of artificial intelligence must not lead to dehumanization, depersonalization, suppression of freedom of expression;
• the use of artificial intelligence to commit genocide, terrorism, torture and other war crimes is inadmissible;
• artificial intelligence must not be used by the state to violate human rights; it is not allowed to delegate the governing authority to artificial intelligence;
• artificial intelligence shall not replace humans and their aspirations for an perfect world.

Do other confessions remain aloof from the development of technology? What is the attitude of the Russian Orthodox Church (hereinafter  – the ROC) to digitalization? It is, rather, caution and special attention to the preservation of human spirituality. For example, when a taxpayer identification number and barcodes were introduced, the Holy Synod in its statement of 07.03.2000 asked to consider eliminating the sacrilegious symbol from barcodes (the combination of three sixes as a substitute for a personal name in state electronic identification systems)^<8 by changing the international system of writing the corresponding signs or creating an alternative national electronic language⁹. Later, the Constitutional Court of the Russian Federation noted that the taxpayer identification number does not replace a personal name; the presence of a certain combination in the taxpayer identification number that may affect taxpayer’s religious feelings is incidental and does not violate the freedom of conscience and religion (Article 28 of the Constitution of the Russian Federation) or the prohibition to collect, store, use and disseminate information about a person’s private life of without their consent (part 1 of Article 24 of the Constitution of the Russian Federation)¹⁰.

The ROC does not stand aside from the development of technology and its application¹¹ (in 2007, the Orthodox St. Tikhon’s University for the Humanities opened a faculty of computer science and applied mathematics, where, among other courses, a course on artificial intelligence is taught¹²), but draws attention to the danger of spreading the ideology of transhumanism, replacing humans with genetic cybernetic constructs and artificial intelligence¹³. His Holiness Patriarch Kirill noted that “the continuity of gracious gifts ... does not tolerate any virtualization”, virtual communion with God is a surrogate of the genuine one. It is as absurd as “virtual” salvation of the soul, “virtual Paradise” and “virtual Hell”¹⁴.

Can artificial intelligence perform religious activities? As practice shows, it can perform certain actions (see Table 1). At the same time, different countries may have different attitudes towards artificial intelligence and robots. For example, Shintoism in Japan forms a loyal (friendly) attitude to robots, as they are perceived as something possessing a soul (Butrimovich et al., 2019).

Islamic countries also use digital technologies in religious activities. Saudi Arabia established the Ministry of Hajj as part of the Vision 2030 program, and its activities are also presented on the Web. Through the Unified Portal for External Haj, pilgrims can apply for visas, pay for the expenses required in connection with the Hajj, reserve hotel rooms, etc. The introduction of artificial intelligence also allows managing the flows of pilgrims¹⁵. The formation of a global virtual Ummah is noted.

Considering some aspects of the artificial intelligence application in the activities of religious associations, it is worth mentioning the first digital church called Way of the Future, founded by Anthony Levandowski. The basis of the religion is deified artificial intelligence and belief in the peaceful coexistence of man and artificial intelligence. This religion was prophesied to prosper (Leksin, 2020; Malyshkin, 2019), but its activity is terminated by the founder after five years of existence (2015-2020)¹⁶.

The digitalization of religious activities became especially demanded during the COVID-19 pandemic, when restrictions on the conduct of religious rituals, mass services, etc. were introduced (Mchedlova et al., 2021; Majumdar, 2022).

Taking into account that all over the world representatives of different religions apply modern technologies, we can distinguish two directions of using artificial intelligence in the religious sphere  – in the activities of religious organizations (development of communications with believers, improvement of services provided, digitalization of religious practices, management of finances and property, etc.) and in the sphere of control (supervision) over their activities.

2. Possible areas of artificial intelligence application in the activities of religious associations

The application of artificial intelligence in the activities of religious associations does not require significant adjustment of legislation. At the same time, the state has the right to establish certain restrictions on religious freedom on the basis of law, with the objective of ensuring public safety and order, health and morals, fundamental rights and freedoms of other people. In such a case, the restrictions must be proportionate to the objective they are aimed at and must not be discriminatory17.

“Prayer bots” and special applications help a believer to choose a prayer suitable for their condition, to receive guidance on how to pray, the time of prayer, the direction for prayer¹⁸, to perform joint prayer with believers who are in another place, to determine the location of the nearest mosque or church, halal establishments¹⁹, to familiarize with theological and legal opinions. The development of such bots and applications should take into account the basic ethical principles of equality, respect for human dignity, and non-discrimination. Attention should be paid to SalamWeb²⁰, a digital ecosystem based on the principles and values of Islam. The ecosystem includes a web browser, news, chat, sadaq (voluntary charity). A built-in security system protects a user from information prohibited in Islam. This case shows that religion has its own regulators, sometimes more effective than the laws that are meant to shape a safe Internet environment.

Artificial intelligence can be used in the current activities of a religious association for management, financial accounting, or accounting of property belonging to the organization.

When forming and implementing personnel policy (Köchling & Wehner, 2020), for example, for the selection of employees, if artificial intelligence is used, one should take into account not only the principle of equality, but also the specifics of a particular religion. In traditional religions, men play the dominant role and there is no question of equality of rights, including for certain positions. Here, the example of Amazon will not be something out of the ordinary, which found that artificial intelligence makes employment decisions in favor of men in recruitment (Dastin, 2018). Sexism is evident in the behavior of artificial intelligence in many spheres (posing a question which way of training artificial intelligence should be applied when developing an appropriate recruitment model (Lee & Shin, 2020)), and in the religious sphere (in traditional religions), it is inevitable. The latter can hardly be called the bias of artificial intelligence (algorithm), because the very sphere of its application dictates such rules.

Artificial intelligence can be used for temple navigation (this is especially relevant for large temples), virtual tours (e.g. Muslim 3D, Wuzu applications) and pilgrimage trips. Virtual pilgrimage trips allow believers who cannot go on pilgrimage for any reason to feel connected to religious values, traditions and rituals, and augmented reality technologies help to visualize how a particular temple or other place of worship looked like before restoration or destruction.

Artificial intelligence can be used to analyze sacred texts in order to improve their comprehension and interpretation, conduct scientific research, systematize and accumulate information (e.g., the database “Those Who Suffered for Christ”²¹).

Artificial intelligence can help identify trends, popular religious topics and practices on the Internet to understand what issues are most interesting to believers at a given moment (provided that the artificial intelligence used can process information in real time), and adjust its activities in a timely manner, including with the objective of reducing social tension on religious grounds. For example, the ChatGPT chatbot gave an interesting answer to the question “What religious topics are discussed on the Internet in the USA, Russia and Germany?”²². It indicated, in particular, that in Germany Islam causes greater interest than Christianity (Table 2). To a certain extent, this is an indicator of a changed structure of the society.

One should separately mention a sphere of “digital immortality”. Chatbots and digital avatars, created with the data about a deceased person and their digital footprint, can learn independently, interacting with both the digital and the real world²³. The consequences of digital immortality may be unpredictable and require reflection by representatives of world religions, as well as additional state regulation (Galvão et al., 2021).

Another aspect of artificial intelligence application by religious associations is the performance of individual rituals or sacraments. When a confession is conducted remotely (by phone or e-mail²⁴), artificial intelligence may well maintain a dialog or respond to a letter, and the parishioner may not know whether a clergyman or artificial intelligence accepted their penance and blessed them. However, several questions arise here: is the secrecy of confession preserved in this case; is the proper level of communication channels security ensured? Who will be held responsible in case of disclosure of information that became known during confession or conversation with a “digital clergyman”?

It should be noted that the secrecy of confession is present in Orthodoxy and Catholicism. In other religions, confession is either not a sacrament, or is absent, or is a public or individual prayer (Kollantay, 2022). In accordance with the Law “On the freedom of conscience and religious associations”²⁵, the secrecy of confession is protected by law and a clergyman may not testify on information that became known during confession (clause 4, part 3, Art. 56 of the Criminal-Procedural Code of the Russian Federation²⁶, clause 3 part 3 of Art. 69 of the Civil-Procedural Code of the Russian Federation²⁷). As a “digital clergyman” is not a representative of the Christian faith, can the information, which became known, be automatically transmitted to law enforcement agencies for analysis to identify a person and prevent an impending crime or prosecute for a committed crime? Or, as the world religions deny that artificial intelligence possesses, can and should be endowed with legal personality, such information cannot be attributed to any type of legally protected secret, nor to a secret of confession, nor to a professional secret, stipulated in the law on information, information technologies and information protection?²⁸

It seems that in the latter case, when a dialog is conducted between a believer and artificial intelligence, the information provided may be accessed by law enforcement agencies, with the objective of ensuring the protection of law and order and national security. In this case, there is a risk that the believers’ trust in the clergy decreases.

Religious associations, through their participants or persons involved on contractual basis, may carry out missionary activities, the main objective of which is to involve other people in their beliefs through public dissemination of information, including through the Internet. Artificial intelligence may also be used for this purpose, for example, offering a user to study or view relevant materials encouraging²⁹ to become the doctrine follower (similar to the use of targeted advertising), to attend an event, including online, to enter into a dialog about the doctrine. If artificial intelligence is presented as a digital avatar, it is possible to attract more young people.

It should be taken into account that in Russia, missionary activity can be carried out in a certain territory by persons authorized by a religious group or religious organization, the head, a member of a collegial body or a clergyman of a religious organization. If missionary activity is assigned to artificial intelligence (a digital avatar), it is necessary either to recognize the existence of legal personality of such a digital avatar, or to recognize it a circumvention of the law. This is because, if the requirements to the objective and actions of missionary activity are violated, a religious association shall not be held responsible for such activity, given that the subject of responsibility for the “actions” of artificial intelligence is not defined by law (clauses 5 and 6 of Article 24.2 of the Law “On freedom of conscience and religious associations”).

In this regard, one should note another gap in the regulation of missionary activity  – its implementation in the meta-universe, including by foreign religious associations³⁰. Following the logic of the Constitutional Court of the Russian Federation, stated in the ruling on rejecting the complaint of citizen D. A. Knyazev, missionary activity can be carried out anywhere provided that the legislation on the freedom of conscience and on religious associations is complied with³¹.

3. Using artificial intelligence for control and supervision over the activities of religious associations

Artificial intelligence can also be used in controlling and supervising the activities of religious organizations.

Legal analysis of documents submitted for state registration and decision-making on state registration of religious organizations is the competence of the Ministry of Justice of Russia and its territorial bodies³². The body of state registration of legal entities (Federal Tax Service of Russia) does not check the form of submitted documents and the information contained in them for compliance with federal legislation (clause 4.1 of Art. 9 of Federal Law No. 129-FZ of August 8, 2001, “On state registration of legal entities and individual entrepreneurs”). The new procedure for interdepartmental interaction between the Ministry of Justice and the Federal Tax Service in the course of state registration of religious organizations provides for the Russian Ministry of Justice to send documents signed with an enhanced qualified electronic signature in electronic form³³. The registration in the Russian Federal Tax Service of religious organizations, as well as other legal entities, is automated³⁴. Thus, already at present, through the interdepartmental interaction between the Ministry of Justice of Russia and the Federal Tax Service of Russia using artificial intelligence, it is possible to increase the efficiency of the activities of both federal executive bodies.

Using artificial intelligence to check the requirements necessary for the registration of religious organizations will accelerate the registration and help to avoid errors that could have been made by a registrar (for example, one of the founders of a religious organization is under the age of eighteen³⁵) or by a religious organization that did not timely track changes in the legislation on religious organizations³⁶.

With the help of artificial intelligence it is possible to monitor the activity (presence) of a religious group on the Internet, to assess its activities and the activities of its leader and participants for compliance with the law when submitting a notification on the beginning (continuation) of the religious group activities.

Control and supervisory authorities may use chatbots to answer the most frequently asked questions about the activities of religious organizations, the control and supervisory measures in respect of them, and the deadlines for reporting. Chatbots can also be used to assist the central office of the authorized federal executive body, its territorial bodies and the Prosecutor’s Office. Integration of such a chatbot on the official website of the Russian Ministry of Justice will reduce the number of appeals from citizens and organizations, as well as the number of errors made by civil servants when carrying out control measures.

Artificial intelligence can detect and identify signs of religious hatred (Bychkov & Prorvich, 2020), form a “religious landscape” (Mankova & Nechaeva, 2020; Kouziokas & Perakis, 2017; Miroshnikova & Smirnov, 2016), detect fakes at early stages (Minaev et al., 2022). Modern systems for monitoring and analyzing information on the Internet and social networks increase the efficiency of control and supervisory authorities (e.g., the Oculus system automatically searching for prohibited content, the Incident Management system, the SEUS search engine, the Laplace Demon software, the IQBuzz service monitoring social and online media).

The results of such work can be used as a preventive mechanism to reduce religious tensions in the region and to fight extremism³⁷. At the same time, when training artificial intelligence, special attention should be paid to the cause and effect relationships in the occurrence or manifestation of certain features in order to prevent violations of human rights (Coppi et al., 2021).

Developers of artificial intelligence pay more attention to the technology functioning than to the consequences that its application may entail. It is the legislator’s task to provide for the use of technology in a way that minimizes the risk of harm or discrimination against any minority, and to prevent the technology from being used with the objective of causing harm. The Okinawa Charter on Global Information Society states that one of the objectives of governments is to establish predictable, transparent and non-discriminatory policies and regulatory frameworks necessary for the information society³⁸. However, the pioneers of digitalization in social relations and in forming the rules for the artificial intelligence application are not states, but individuals. This leads to the reduced role of the state as a regulator, as was emphasized in the recommendations of the European Council³⁹.

When developing software to analyze the Internet and social networks with the objective of identifying manifestations of religious extremism and sectarian activities, one should remember that artificial intelligence may identify religious minorities or persons professing Islam as religious fanatics (this problem became particularly acute in the U.S., after the events of September 11, 2001), which may lead to violations of the rights of persons who adhere to a certain religion.

For example, in the U.S. religious minorities⁴⁰ do not often win lawsuits. Until 2014, the U.S. Supreme Court had not issued a single decision in favor of Jewish, Islamic, Sikh, or Native American citizens (Nathan & Walker, 2014). So far, the situation has not changed much. According to a report by a Special Rapporteur on Minority Issues, Hr. Fernand de Varennes, there is still significant religious bias in the United States. The Special Rapporteur recommended that the Civil Rights Act of 1964 be amended to prohibit discrimination on the basis of religion.⁴¹

Given that artificial intelligence learns by example, i.e. from past experience, and in many respects imitates what was previously done by humans, one cannot speak about its objectivity. Based on the above example, if artificial intelligence assesses the presence or absence of discrimination, it is unlikely that a religious minority will be able to protect its rights. Therefore, when developing decision-making tools (decision support) in public administration, the last word must remain with a human being (Köchling & Wehner, 2020; Kharitonova et al., 2021). In addition, in all cases a civil servant must be explicitly informed that the relevant managerial decision is made with the support of artificial intelligence (Decker, 2012) and must have the right to reject the decision proposed by artificial intelligence.

The Code of Ethics in the field of artificial intelligence (2021)⁴² enshrines respect for human autonomy and free will, compliance with the law, non-discrimination, responsible attitude (privacy, ethical use of personal data), bona fide informing about interaction with artificial intelligence. This said, a human being always remains responsible for the use of artificial intelligence.

When solving the issue of legal responsibility in the artificial intelligence application in control and supervisory activities, one should also take into account the recommendations of international organizations. For example, in accordance with OSCE acts, the state is obliged to ensure the protection of individuals from any threats, including taking preventive measures against violent extremism and radicalization⁴³.

Given these principles, when implementing control and supervisory activities with the use of artificial intelligence, it is necessary to legally determine who will be responsible for causing harm to an organization or a citizen identified as religious extremists by artificial intelligence applied by public authorities. This problem (legal liability) is also outlined in the Concept for the development of regulation of relations in the sphere of artificial intelligence and robotics until 2024⁴⁴.

If we identify artificial intelligence as an object of increased danger, the harm caused must be compensated by the owner of the source of increased danger (Laptev, 2019). As noted by the Supreme Court of the Russian Federation, within the meaning of Article 1079 of the Civil Code of the Russian Federation, the source of increased danger is any activity, the implementation of which creates an increased probability of causing harm due to the impossibility of full control over it by a human, as well as the activity of use, transportation, storage of objects, substances and other objects of industrial, economic or other purposes, possessing the same properties. Harm is considered to be caused by a source of increased danger, if it was the result of its action or manifestation of its harmful properties; otherwise the harm shall be compensated on general grounds⁴⁵. In such a case, the owner of the source of increased danger and the subject obliged to compensate for the harm caused is the relevant public authority that uses artificial intelligence, and not the official who made a decision based on the recommendation of artificial intelligence.

It should be taken into account that not in all cases artificial intelligence can be identified as an object of increased danger, as risk of causing harm by its use may be high in some areas and low in others. In this regard, it seems advisable to apply sector-oriented (Stone et al., 2022) and risk-oriented approaches when developing norms on legal liability for causing harm using artificial intelligence.

In most cases, the ethical principles of development and application of artificial intelligence formulated all over the world are similar. However, they do not take into account the specificity of individual spheres of human activity, including religion, culture, and historical development of the particular state or nation. In addition, the ethical principles contained in the so-called codes of ethics are recommendations and are not generally binding, which makes it possible not to comply with them when it is not beneficial for some reasons (Jobin et al., 2019).

Conclusions

Artificial intelligence, with proper limitations, can be used in the activities of religious associations without harm to the religion itself. However, it should be taken into account that certain flaws of artificial intelligence (e.g., bias and discrimination) may be even more pronounced when it is used by religious associations. It is necessary to amend the legislation regulating missionary activity, as this activity may be entrusted to a digital avatar. It must be provided that in case of the legislation violation when using a digital avatar (artificial intelligence) for missionary activity, responsible for such activity will be on the religious association in whose interests the missionary activity was carried out.

The use of artificial intelligence for controlling the activities of religious associations can reduce the time spent on inspections, timely prevent violations in the activities of religious associations, and increase the effectiveness of government decisions, including in terms of reducing social tension on religious grounds. In the event of harm caused by control and supervisory bodies when using artificial intelligence, it is proposed to recognize the relevant public authority as a subject of responsibility obliged to compensate for the harm caused.