Artificial Intelligence (AI) has the ability to complete complex tasks in a matter of seconds and makes available democratized and easily accessible knowledge for many people. It also is an excellent tool for overcoming any creative block. On the other hand, it can complete complex tasks without generating any real learning. There are also complex issues of copyright infringement that need to carefully be taken into consideration, and there is the real possibility that content created with the help of AI may be incorrect. These are just a few of the benefits and drawbacks commonly raised about new generative AI tools in conversations or analyses on the internet. These innovations are even pointed out by many experts as the beginning of a new cycle in the behavioral and technological development of the world.¹

Every year, the largest technology companies strive to dominate the global market by releasing new devices and tools. None of these releases are everlasting; they will all be replaced by newer versions. However, some of them have the potential of significantly impacting or altering human behaviors or even influencing human brain development.² These inventions assist human needs, but also awaken previously nonexistent desires within us. They are tools that were previously inaccessible to the majority of the population. This was the case with the launching of laptops, smartphones, and social networks. We cannot imagine our world today without them.

The fact is that AI is revolutionizing everything in the technological world. From a refrigerator with a display to a microphone, which until recently was considered “naive,” AI is collecting data and using it to map our behavior and enhance their services.

The broad impact of this revolution raises the question of how AI can be used for theology and what its implications are for Christian ethics.

New and Old Together

AI, as a whole, is not a recent invention or tool. Its origins can be traced back to the 1940s with Alan Turing’s work, which helped break the German Enigma encrypting machine’s code during World War II. The foundational concepts of AI were further explored with the publication of Turing’s article “Computing Machinery and Intelligence” in 1950. This is considered by many as the beginning of AI. However, Turing never used the term “artificial intelligence” himself. The term would be used for the first time in 1956 at a conference called “The Dartmouth Summer Research Project on Artificial Intelligence” in New Hampshire.³ Since then, this field has experienced many changes.

The great hope was that AI would revolutionize various segments of technology in a short period of time. This perspective led to significant investments from companies that saw AI as a hopeful solution to problems and a means to create advantages against their competitors. Herbert Simon, a political scientist who impacted such areas as economics and psychology, exemplified the hype around AI in the 1960s when he said in 1965, “Machines will be capable, within 20 years, of doing any work a man can do.”⁴ This statement still does not fully apply today, but we are moving in this direction.

Despite some advances in the early 1980s, AI suffered something known as the “AI winter” from the 1970s to the 1990s. Major companies did not see the expected return from their significant investments in the field. In the 1990s, hope arose when IBM’s machine called Deep Blue defeated Garry Kasparov, then world chess champion, in a chess match in 1997. The machine had already lost a match the previous year, but this evolution of capacity in the space of just one year rekindled the mood for a more optimistic future for AI.

Since then, this area has grown significantly and has become present in our lives in a way that we often do not realize. Our cell phone has hundreds of AI applications working in various activities. As of a few years ago, every social network, spell checker, GPS app, streaming platform, and even search engines used AI. For example, when you type a word in Google and search suggestions appear to complete your sentence, nothing is by chance. These suggestions are provided by an AI that analyzes your age, search history, trending topics, and possible preferences. After mapping them, Google delivers possibilities that we are most likely looking for. This AI is called Google Suggest.

But if AI has existed for over seventy years, and if we have been in contact with it for decades, why is this theme on the rise again? Why has it become a global concern? The most plausible explanation is that we are entering now into a new phase in which AIs are not only used by large companies to map our behaviors, collect our personal data, or improve our app experiences, but can now also be used by users themselves in creating content for themselves.⁵ This is known as generative AI. Today we can open incredible technological tools and create images, texts, books, presentations, and many other things instantly. People without coding experience or technical knowledge can create a mobile app quickly, simply by writing what they want and following the instructions.⁶ In a few seconds and without any artistic knowledge, we can create images that previously could only be made by someone with a special talent and willingness to devote a significant amount of time to the project. We can create extensive texts on certain topics without even mastering the topic and without spending hardly any time. Unlike before, we would not just be copying ready-made content from the internet but generating something entirely new. This raises some important ethical questions that are connected with generative AI.

Morality in AI

With such convenient power at our fingertips and the advancement of conversational AIs such as ChatGPT, Google’s Bard, and Microsoft’s Bing, ethical questions become unavoidable. Who is the “owner” of the generated content? If someone creates a complete article using an AI, for example, could that person be considered the author of the article, even if it was written by AI? What Christian principles should we uphold when using these tools? As Christians, should we have a favorable or unfavorable stance towards these new technologies? What are the main risks of using AI? And finally, is it possible to use AI in preaching and sharing the gospel? These are just some of the questions frequently found on the internet. Many of them still lack a consensus answer.

There are many other technical or ethical questions that we will not address in this article, such as the issues related to unemployment and its global economic impact. Many of these questions, in fact, still lack satisfactory answers.

The moral and ethical discussion in the field of AI is not a recent phenomenon. This concern has permeated the entire history of technology development. One of the key points of discussion is the moment when machines become as capable in all areas of knowledge and human skills as humans themselves, which is known as Artificial General Intelligence (AGI). In this phase, the machine would not just learn specific aspects as it does today, but would be capable of developing in any area. This concept is not new either; it was developed by Alan Turing himself in the aforementioned 1950 article and became known as the Turing Test.⁷

But it does not stop there. AI could supposedly reach singularity in the next phase, surpassing humans in all aspects. This is known as Artificial Superintelligence (ASI).⁸ In this phase, humans would no longer have control over the machine, and the machine would have complete autonomy to make decisions and perform tasks.⁹ In both phases, the machine would not only execute functions as it does today, but would also have an understanding of why it performs them—something that does not yet exist. For many, humans would no longer be the most intelligent and capable beings on the planet. Scientists have not reached a consensus on when this phase would begin, with predictions ranging from five to thirty years, and some even believe that this moment will never arrive.

If AI were to someday reach that level of development, several other delicate aspects would come into play. Some of these were addressed in an open letter published on March 22, 2023, by the Future of Life Institute (FOLI) and signed by many thousands of CEOs, researchers, and individuals connected to the technology sector. The letter, which made history in the development of AI, suggested a six-month pause in its advancement to discuss the consequences of its use. While the request from the letter obviously was not heeded by the scientific community, the discussion about its impact has received greater attention. “AI systems with human-competitive intelligence can pose profound risks to society and humanity,”¹⁰ the letter begins. What are some of these risks? Unemployment, rapid extinction of professions, more catastrophic conflicts, and other consequences that are still unknown.

However, we should not think that concerns are only tied to ASIs. The generative tools we use today also have their own concerns. These tools are trained on a learning model known as Deep Learning, which requires massive amounts of content for effective results, as the model learns primarily through examples. As a result, these tools create responses by cross-referencing data and information. Generative tools have implications in many areas of human knowledge, and as a result, discussions about their good and bad usage have extended beyond the technological realm and become concerns across nearly all fields.

Among the main concerns are privacy, both of the content used to train these tools and of the users themselves. Another concern is the issue of deepfakes, because these tools have the capacity to generate hyperrealistic texts, photos, and videos about nonexistent events in a matter of seconds. There are copyright issues and the manipulation of public opinion on sensitive topics. Lastly, we have to consider the implications of these tools on new human and societal behaviors.

AI also significantly impacts the education process. Teachers find themselves facing the challenge of redefining homework, a crucial part of the learning process. A study conducted by the website study.com with one thousand students found that as of January 2023, 89 percent of students reported using ChatGPT to complete homework assignments and over half of the respondents admitted to using it for creating school texts.¹¹ Since ChatGPT allows students to add specific requirements such as an exact word count, the number of pages, or even the literary style, many students have not even read the provided text and simply submitted what AI generated to meet the requirement. The problem is intensified by the inability of many educators to determine whether assignments were entirely done using AI. Surprisingly, the study also showed that almost three-quarters of the interviewed students favored banning ChatGPT from the educational process, indicating that they are also concerned about the consequences.

Johann Neem, a history professor at Western Washington University, pointed out to The Wall Street Journal the issue of using tools like ChatGPT in education: “Just because there is a machine that will help me lift up a dumbbell doesn’t mean my muscles will develop. In the same way, just because there is a machine that can write an essay doesn’t mean my mind will develop.”¹²

Although some teachers and universities are in favor of integrating these tools into the educational process, many universities around the world have restricted the use of conversational AIs like ChatGPT and Bard on their campuses.¹³ Other educators have tried to mitigate the consequences of using these tools through alternative methods, such as handwritten assignments, small reading quizzes, or oral assessments. Some are even moving towards requiring tasks that involve personal critical analyses of lesser-known works, with questions that make it inevitable that the student must read the content and form a specific opinion, such as identifying the drawbacks of the work or drawing connections between the content and what was covered in class. Finally, some teachers have been adding a small declaration at the end of assignments where the students must affirm that they did not use any AI or similar means to complete the proposed activity.

While it may still be early to determine which methods will be the most effective, these new challenges in the field of education underscore the importance that teachers, parents, and all those involved in education must place on motivating students and conveying the true significance of the learning process (including extracurricular activities). It is not wise to rely solely on the personal motivation and maturity of students to prevent them from misusing such tools. More than ever, instructors must focus on shaping the character and integrity of students. They should educate students to have a full awareness that generative tools do not replace the benefits of the learning process and the relevance of knowledge. Just because I am driving a car faster than riding a bicycle does not mean I am getting more exercise from the car. Learning involves more than mere memorization; it encompasses intellectual and character development and critical thinking.

In the academic sphere, master’s and doctoral students have also been using ChatGPT in their thesis work. However, the tool poses a problem for unrestrained usage. As seen earlier, generative AI creates responses using a lot of content available on the internet. However, often the user cannot verify the source of the material used in the research.¹⁴ They cannot even mention or add it to their bibliography because due to the generative nature of the intelligence, the provided response will be different every time it is asked, even if it is the same question.

And So What

Given that AI systems are but tools, one should not assume that they are inherently negative or positive. Just as with social networks, communication tools, and applications, it is the manner of use that dictates whether they shall be deemed beneficial or harmful.

As enthusiastic or concerned as many may be about these new tools, it is the duty of every pastor and religious leader to be aware of the potential as well as the risks of such technology. They need to raise an awareness among their members of the risks and problems that the misuse of these tools can bring, while also highlighting the benefits of using them correctly. Churches can address this topic in a creative manner, discussing it with young people in small groups or informal settings. Schools and colleges can utilize chapels or guest lectures to engage students with this topic. We should encourage our youth to become role models in the use and development of these technologies, especially for noble purposes. Many universities have already taken steps to teach students how to use these tools, but it is also important to prioritize ethical and moral discussions about their use.

Exercising common sense during the development process of new tools is essential, especially in the context of recent initiatives emerging in the evangelistic field of the Seventh-day Adventist Church linked to AI. In this regard, Christian ethics play a crucial role, demanding that professionals involved in the design and creation of tools consider moral and spiritual principles in their approaches. We can strive to humanize the experience and interaction of users with AI systems while ensuring a clear distinction between the identities of the communicating entities—whether they are machines or human beings. It is crucial to acknowledge that AI tools are subject to errors, as they operate based on probabilities. Consequently, content should never be automatically shared as an official church position without making sure that the response is correct and in harmony with what the Seventh-day Adventist Church officially affirms.

The way AIs, such as ChatGPT and Bard, train their models using content available on the internet underscores the importance for the church to consistently produce high-quality articles and content, making it available on our websites. This is significant because various themes of biblical interpretation and subjects that are unique to Adventist theology are not readily available from an Adventist perspective on the internet. Sometimes, it is easier to find critical analyses produced by theologians from other denominations regarding Adventist interpretations than it is to find Adventist interpretations themselves. Regular postings of high quality articles on our websites will also aid in ranking higher those sites on Google. This content will increasingly be found not only by visitors to these sites, but also through other means. This is why the church and their members need to be active in digital ministry and need to produce and post quality theological articles on their websites (especially on websites that do not carry the church’s name in the domain).

AI can also support the church in the communication of information and in reaching larger numbers of people. We should not avoid or ignore AI simply because there are numerous examples of misuse. Instead, it is important to understand that the responsibility for an ethical and proper use of these technologies falls on us, as creators and users. We must comprehend that new technologies lead to new behaviors, and new behaviors provide new opportunities to spread the gospel to others. Here is an illustration of how AI was used in the Seventh-day Adventist Church in South America.

Through an AI called “Esperança” (“hope,” in English), Novo Tempo (the Hope Channel in the South American Division territory) has engaged with nearly three hundred thousand people in studying the Bible over the past four years. It was the first Adventist project using AI for evangelistic purposes.¹⁵ People from over a hundred different countries have gained a better understanding of biblical truths through a robot that responds to common questions and interacts on social media with individuals in Portuguese, Spanish, and English. Since May 2019, over fourteen thousand people have decided to commit their lives to Christ after studying in this unique way. Individuals from countries and regions that otherwise would be extremely difficult and dangerous for missionaries to reach have been reached because they only require the basic prerequisite for studying: a social network and interest.

Last year (2022), Esperança began answering biblical questions using AI. It utilizes biblical texts and audio-visual material produced by the Seventh-day Adventist Church, and uses more than six thousand articles and excerpts from the works of Ellen G. White to formulate its responses and provide additional content to encourage users to continue studying. This is a small example of how AI can and has been used in sharing the gospel.

One might question whether we are “robotizing” the human mission, which should be the role of church members. However, it is important to consider that such tools are not meant to replace the personal mission assigned to each Christian (Matthew 28:18–20), but rather to amplify it. We should also not forget that behind these tools are people with their gifts and talents working to foster and advance the mission of the church. The AI tool is merely the means that is used in this endeavor. During the process of studying with interested individuals, the participation of church members is crucial in welcoming and providing spiritual support for these potential new converts. Just as the church has utilized various other technological means to reach people, these tools cannot replace the personal interaction and contact but serve as catalysts to reach even more individuals for the kingdom of Christ.

Every pastor, researcher, or educator should have at least a basic understanding of AI tools like ChatGPT. If you’re not familiar with the tool, consider the possibility of taking a basic course. You can use it effectively only if you submit the right questions to AI. The more specific your search questions are the more effective AI will be in assisting in research, suggesting sermon themes, and even aiding in the development of Bible studies. AI can correct written texts, improve the language, or translate text or spoken messages with high accuracy. While the answers may not always be 100 percent precise (especially when texts from Ellen White are translated), these tools generally have a high level of accuracy, and if used appropriately, they can save a lot of effort and energy.

Give It a Try

As previously discussed, there is a strong debate regarding the potential of AI to replace jobs and professions in the future. However, it is a consensus that those who fail to adapt and capitalize on the benefits of these technologies will see their productivity decline compared to those who have adapted. To illustrate it, consider the difference in productivity between a professional food delivery with a horse-drawn cart and another using a car. Though they might perform similar tasks, the time to achieve similar results will be drastically different.

Every pastor, researcher, or educator ought to possess at least a foundational understanding of AI tools like ChatGPT. If unfamiliar with this tool, one might contemplate enrolling in an introductory course. It is hard to get good answers without knowing how to ask the right questions to these tools (an example will be provided later). Such technologies can significantly facilitate research, propose sermon themes, and even assist in creating Bible studies. They can correct both structure and spelling of texts, modify language to be more accessible or more academic, or translate with good accuracy. Educators can employ this tool to easily correct academic papers, for instance, by requesting the tool to critically evaluate each assignment based on the Adventist perspective on the subject or by feeding it with the content covered in class.

Another outstanding capability of generative AI is to condense texts into their key points or to underscore the most salient aspects. Consider summarizing several pages of text into just one or two paragraphs. AI can also offer creative ideas for thematic divisions when crafting new materials. For example, if you are planning to write a fifteen-chapter book on soteriology, ChatGPT can give you ideas for what to cover in each chapter, keeping your target audience in mind.

For educators, AIs can help share knowledge by offering ways to support students with special needs or to teach a certain subject. They can make pictures, give teaching examples, or even come up with deep thoughts for big discussions. A teacher could use the tool to make a test, just putting the content in and getting the questions out. With its easy-to-use setup, lots of customization, and a ton of data, AI can help anyone doing research on any topic from start to finish. While answers might not always hit the mark (especially concerning Ellen White), these tools are usually remarkably accurate and, when used wisely, can save a lot of time and energy.

AIs can also help find solutions to problems in specific areas, evaluate the effectiveness of existing plans, or even refine ideas that are not fully matured.¹⁶ It is essential to use these new technologies wisely in our personal research. The results provided should only be an initial step in the research process, and not treated as the final word. We should not ignore the fact that, being a generative tool, in many cases the responses provided by ChatGPT or any other generative platform may not be 100 percent accurate. The availability of content on the requested topic on the internet will likely affect the precision.

Therefore, the work of researching external sources and confirming results through other means becomes crucial. Furthermore, a correct answer from a non-denominational tool such as this may not necessarily reflect our specific understanding of the topic. In the research field, Google’s Bard, which is still in its testing phase, and Microsoft’s Bing have an advantage over ChatGPT (at least up to version 4) because they provide some of the references used in creating the response. Therefore, these tools are not only interesting for accessing a possibly different argumentation and approach, but also for obtaining some references and bibliographic information.¹⁷ Bard also has another advantage of being connected to the internet, which can assist in researching recent topics. However, it is always interesting to use more than one generative tool, as the answers and approaches can vary.

On the other hand, ChatGPT (particularly its version 4) has been achieving notably superior results compared to Bard and Bing currently. However, to get the most out of this tool, it is important to know how to ask the right prompts (commands). A simple tip is not merely to ask a direct question, but also to provide a background, context, or specific nuances you seek in the answer. Such details will enhance the AI’s precision. If the response does not meet your expectations, changing even a single word in the question might produce a quite different and possibly more fitting result. Yet, this is not always necessary; expressing your dissatisfaction or further detailing your request can deliver a more precise answer. Concerning context and details, imagine you wish to draft a brief Bible study on the gift of tongues for classroom use. If you simply request verses addressing the gift of tongues, you’ll obtain a basic list. If you quickly ask for an analysis of these verses, the response might contradict the traditional Adventist perspective. But for your study’s objective, when posing your question, it is essential to be more specific. For instance, you might input the following prompt into ChatGPT: “Imagine you are an Adventist theology college instructor. Your assignment is to craft a Bible study with an introduction, five verses from the Pauline epistles about the gift of tongues with individual analyses of each, followed by five true-or-false questions of a moderate-to-high level of difficulty. Conclude the study with a Christocentric and missiological reflection, linking it to the thirteenth Adventist belief: The Remnant and Its Mission.” In this case, you could be more specific with your request, like asking to include Greek words, turning the study into a poem, or even adding fun riddles in the questions. The possibilities are endless.

Many other AI tools are specifically designed for research and can be very useful in academia or education in general. For example, Consensus is a great tool for finding relevant articles in a particular area of interest. When conducting research, the tool provides a summary, making it easier to assess relevance, provides some basic information, and offers a link to the complete article. Research Rabbit aids in the process of creating bibliographies, tracking citations, and generating summaries of papers. Time-consuming processes in the production of scientific articles can be streamlined through intelligent tools.¹⁸ To develop slide presentations, some tools are available such as Gamma, beautiful.ai, tome.app, and even PowerPoint’s own Copilot 365. Most of these tools are paid and require a bit of practice to achieve meaningful results.¹⁹ However, if you are not technologically proficient, just being familiar with ChatGPT or Google Bard will significantly improve your research process.

What can AIs not do? They will never replace the need for human interaction, the importance of collective Scripture study, biblical authority and centrality, and the necessity to open the Bible and seek the true meaning and significance of each passage, without external interpretations. In spiritual terms, AI tools should not overshadow our intrinsic need to meditate and reflect on biblical topics or texts under the guidance of the Holy Spirit when preparing some material. Although these tools can craft entire sermons and articles, their primary role should be supportive in their preparation, ensuring that the connection and communion with God remain as central elements of the whole process. Even if AI can provide insightful answers based on extensive databases, it lacks the ability to discern the individual challenges and needs of your congregation or classroom. It would not know the peculiarities of each individual and, as a result, what would be the most effective approach. At best, it can only offer suggestions.

It is important to emphasize that we are still in the early stages of the AI revolution, and new tools will be launched in the future. It is crucial to have an evangelistic-missionary perspective when considering the new possibilities that will arise with these new AI tools. How can we use this tool to better prepare ourselves for preaching the gospel? Is there a way this tool can be useful in our institutional work? Will this tool make preaching or carrying out the mission easier for you? How can you integrate this tool into a missionary initiative? These are some of the questions that need to be asked.

If God has used radio (AWR), television (Hope Channel), social media, and so many other technological means to reach people, why should we not be able to also use AI? I am convinced He will use them as well.

As you explore the potential of AI tools in your ministry and mission, consider how they can enhance your abilities, expand your reach, and extend the message of the gospel to new audiences. Just as technology has been a powerful tool for spreading the Word of God, AI can be harnessed to further advance the mission of the church and impact lives. Keep in mind the ethical and moral considerations, while also being open to innovative ways of sharing the message of hope and salvation.