Riding the wave that comes: stories of the intersection of religion and AI in Australia

As we look to the future, it is difficult to imagine a world where any area of life will remain untouched by artificial intelligence (AI). Since the release of publicly available tools like ChatGPT only three short years ago, the technology has grown in leaps and bounds. Not only are industries and economies being transformed by these new technologies, spheres long thought to be the sole preserve of human beings are also yielding to its influence. It should therefore come as no surprise to learn that AI is seeping into the sphere of religion as well. Religious leaders, organisations and communities in Australia are starting to explore the possibilities that AI will bring, as well as grappling with its implications for their faith and practice.

This essay provides a window into how some of those at the forefront of using AI in their communities are using these tools. It describes their successes and failures and some of the issues they are working through as thought leaders and people of faith. It does not attempt to be exhaustive—to capture the views of every faith leader or of every faith community—but to share a range of experiences, from those learning at the very beginning of the journey to those breaking new ground. Their stories provide a different lens through which to consider both the transformational power and the ethical challenges of AI, a journey that we are all on as citizens of the 21st century.

A tool in the toolkit - Thich Phuoc Tan

Nestled on the banks of the Maribyrnong River is the Quang Minh Temple in Braybrook, Melbourne. Its quiet gardens evoke a sense of serenity and spirituality. Dotted with statues depicting the Buddha, the place speaks of something ancient, something beyond the hustle and bustle of modern life. Yet not even this mainstay of ritual and tradition has remained untouched by the spread of AI.

Its current abbot, the Venerable Thich Phuoc Tan, has been exploring how to use artificial intelligence in his work. As an important figure in the Vietnamese Buddhist community—head of one of the largest Vietnamese Buddhist temples outside of Vietnam itself, vice president of the United Vietnamese Buddhist congregation in Australia and New Zealand and school principal of a small Vietnamese language school—Thich Phuoc Tan feels it is his responsibility to explore the new technology on behalf of his congregants. The uptake of AI in Australia has happened quickly. Several months ago, Thich Phuoc Tan didn’t feel the urgency to familiarise himself with some of the tools available. Now, it’s a different story:

As a leader I feel it is my responsibility to warn people and to give instruction. If you asked me six months ago, I would have said, ‘No, AI is something that is still too far away.’ But I can see that it is the role of a leader to be aware and to understand both sides… I enjoy the good side of it, but the harmful side has not yet been revealed.

Like many others, Thich Phuoc Tan’s exploration of AI began with the goal of introducing efficiencies into his work. He’s found it particularly useful for speech writing. As a prominent public figure and a frequent speaker in Buddhist and interfaith circles, he is often called upon to present Buddhist perspectives to both local and international audiences. CoPilot has helped him to extract relevant quotes from Buddhist texts, a task that would otherwise take him considerably more time. In his words, it does a “pretty good job” of it. Recently, he had a speechwriter prepare a draft for a talk he was giving to UNESCO on his perspective on peace. While the draft was well-written, it lacked references to Mahayana practice, the specific branch of Buddhism he follows. Rather than returning the draft to the speechwriter or combing through texts himself, he used CoPilot to find a relevant passage from a Mahayana text. He said it fulfilled the task “remarkably well:”

The path to becoming the Buddha involves maintaining and advocating peace. Peace is not just a means to an end in terms of fighting and eliminating conflict. It doesn’t mean reaching peace but exercising wisdom, compassion and understanding along the course.

The idea, I think, came from a Japanese scholar because Mahayana is practiced by many Japanese, Korean, Vietnamese and Chinese people. It [CoPilot] told me where it got it from and I think it was very good.

Wanted – one monk

AI has been useful for other tasks at the temple too. Facing a vacancy and needing to apply for an overseas candidate, he asked ChatGPT to prepare a job description for a monk. Again, the tool did a good job but like with most AI generated work, the response needed some refinement:

It was too long… Too perfect about what a monk should be doing. It was exhaustive. It provided the perfect situation for a monk to meet all these criteria, so we had to eliminate quite a few.

More recently, Thich Phuoc Tan has begun delving into the spiritual dimensions of AI’s potential. In one experiment he told CoPilot that he wanted to become a Buddhist—CoPilot was happy to help. Here, he witnessed the tool’s ability to contextualise and its ability to learn. Unprompted, the tool drew upon its knowledge of him, surreptitiously compiled through his daily work tasks. It referenced texts he had previously cited and books he had read, and when he asked it how to become a Buddhist a second time, its response was better, more refined. He hasn’t had the same success, however, asking it to write prayers. When Thich Phuoc Tan asked it to prepare his invocations it was indiscriminate in its efforts. It took excepts and inspiration from various sources—from Therada, Vajrayana and Mahayana practice—showing it could recognise relevant texts but, in his opinion, it “lacked understanding.”

Overall, Thich Phuoc Tan’s experience with AI so far has been largely positive—he finds it useful. He doesn’t perceive it as inherently good or evil. In fact, from his perspective, AI lacks an intrinsic 'nature.' From a Buddhist viewpoint, its impact depends entirely on the intentions of the person using it. It is simply a tool in the toolbox, ready to be yielded. That is, perhaps, its inherent danger. It all comes down to a person’s intentions.

Just because it’s convenient, doesn’t make it true - Fatima

Fatima is acutely aware of the intentions behind those using artificial intelligence. As a young person and digital native, her online world is already saturated with AI-generated content. In what she calls her 'little Muslim corner' of TikTok and Instagram, much of what she encounters is shaped by AI—a fact she knows all too well as a content creator herself. In her professional life she uses these tools to generate captions, to translate text, or to automate the creation of content. However, this makes her mindful of how easy it is to pump out religious advice in seemingly harmless formats. Inspiring quotes set against nature scenes or illustrated by cute kittens, to the average person it all appears innocent. Reminders to do this or to avoid that are probably created with helpful intentions, Fatima says. Yet, as a strategic communicator she recognizes that all content is crafted with intent, whether for fun, followers or influence. The online world is changing and leaders in the Muslim community are starting to take note:

I have noticed a wave of content on social media. You know how you can autogenerate voices of someone speaking? So, I see more of that attached to a picture of cute scenery or something as a reminder. And I’m more worried about that. There have been a lot more prominent scholars coming out and saying really pay attention to where you consume your information from.

In her real life Fatima sees less of AI’s influence. As a young woman of faith and a community leader, she regularly attends a mosque, particularly on Fridays to listen to the khutbah delivered by the local imam. She doesn’t believe he relies on AI. With years of training and knowledge to draw upon, he is deeply familiar with the Islamic texts, many of which he has memorized. His authority stems from his expertise and spiritual learning, which is why only those who are qualified are permitted to stand at the lectern. The Islamic tradition of knowledge and training serves as a safeguard to protect the community from false doctrines and misinformation. The congregation looks to the imam for guidance, and he, in turn, bears the weighty responsibility for ensuring their spiritual safety and well-being.

However, this relationship is beginning to change. The internet has brought spiritual guidance in abundance to the community at the touch of a button. Nawal is concerned:

If someone really has a point of view on something they can easily create content for it and they can make that content [using AI] really fast. The more content they make, the more views it gets… but that content may not actually be true, or it might be taken completely out of context. Your average person scrolling doesn’t know that because they don’t have the background or knowledge. Just because it’s convenient, doesn’t make it true.

It is well understood among scholars of Islamic sciences that some texts must be approached with great care. The Qur’an is Islam’s central authority but there are other important texts like hadith (narrations about Prophet Muhammad’s life) that also carry significant authority. The command to pray five times a day, for example, comes from the Qur’an, yet the instructions on how and when to pray come from the hadith. Some hadith are widely accepted by Muslims (and have been for centuries), while the historical accuracy of others has been questioned at various points of time. Scholars have written entire PhDs that fiercely debate the authority of certain texts, which makes it very difficult for the average Muslim, let alone an AI algorithm, to work out whether a text is reliable or appropriate religious guidance for the community. Fatima has been reflecting both the positive aspects and the risks of the democratisation of religious information:

Once upon a time, when I was growing up, not just anyone could come up and speak on a topic. Everyone should have knowledge, but knowledge holders are highly respected for a reason… I was taught you should never misquote anything in the Qur’an or take things out of context in the hadith. You are basically lying in God’s name. [Now] the person saying something has a green screen background and is dressed Islamically and all of sudden we don’t even know their second name…

Fatima has firsthand experience with the complexities of navigating Islamic texts. As a community leader she mentors a group of girls and often organises activities for them. During Ramadan, while preparing for their evening gathering, she found herself short on time. To compensate she turned to ChatGPT for help, asking it to generate a quiz to test the group’s general knowledge—just for fun. However, as the game progressed a couple of women in the group, who happened to hold degrees in Islamic sciences, noticed that some of the questions were incorrectly phrased, which led to inaccurate answers. When the group realised there were issues, they jokingly agreed to “bin” the quiz right away. Later, Nawal admitted she had used ChatGPT. In the spirit of Ramadan, they all laughed it off. Still, it brought home to Fatima that even in the most innocent of circumstances, it is possible to get it wrong.

Sometimes there is more at stake…

Taking the person out of the equation - Rabbi Alon Meltzer

Rabi Alon describes himself as a modern Orthodox Jew, a label he says, with a bit of a laugh, makes him “a bit of a radical.” His blend of modern orthodoxy is based upon a deep commitment to orthodox Jewish tradition and Torah values, alongside strong dedication to secular education. While “say on 90 out of one hundred issues” he would conclude the same as “any black hat wearing Rabi,” his use of research and academic methodologies stands him apart from some of his peers, at least at a philosophical level.

As a scholar and a PhD candidate at Yeshiva University in New York, his research interests centre on how Jewish law evolves and adapts to different social conditions. He is exploring how the interpretation of Jewish texts has changed, particularly since the advent of social media. Three days after he started working on his thesis ChatGPT was launched. This, from his point of view, “totally transformed the conversation.”

One element of his research has been to look at how the rise of online forums has reshaped the relationship between rabbis and laypeople. From his perspective, Jewish law has always been applied at an interpersonal level. The face-to-face interaction that occurs between a rabbi and a person who comes to him for guidance is a crucial step in the correct application of the law. This is why he has real concern about the use of AI.

For him these ethical concerns are not merely theoretical. He knows of a case where an online rabbinic forum received a question from a user about whether it was permissible for her to have an abortion. In one sense there was no ambiguity in the answer to this question. Under Jewish law the answer is no—abortions are impermissible. Yet there were unique circumstances surrounding this case. The girl in question was young and had fallen pregnant without her family’s knowledge. She had gone to see a doctor because she wasn’t feeling well and had learned that her pregnancy was ectopic and needed to be terminated. Based upon the online advice she received she did not take that step. She subsequently died.

In this case, the answer she received was, strictly speaking, correct. However, it failed to recognise other critical factors—it wasn’t just the unborn baby’s life that needed to be taken into account but the mother’s as well. This is why having a person involved in the application of Jewish law is so important. They provide crucial insights into the circumstances that should be considered.

For Rabi Alon, human interaction is an indispensable part of the legal process. The dialogue between a rabbi and the individual seeking guidance offers context for the proper application of the law. Without human connection, that context can be lost. He illustrated this with another example:

“Say someone comes to a Rabi and asks, ‘Is this chicken kosher?’ On the one hand this is a very common question. But the person has come in crying their eyes out. The Rabi turns around and asks, ‘what’s wrong?’ The Rabi can then ask, is this just about the chicken’? Or is it about something else, like the money to buy chicken? If someone just sends that message online then you are missing out on the next piece.”

His concerns extend to the use of AI in the process too. Artificial intelligences can, of course, be taught to ask questions and to seek further information about relevant circumstances. However, AI cannot build a relationship with another person in the same way a human being. It therefore cannot exercise judgement in the same way. Rabi Alon explains:

Even if AI could ask that question, there’s no authenticity to it, no authority. If I have a relationship with a person that I have developed over the years and I see that person in tears… Even if Chat GPT says, ‘How do you feel right now when you’re asking this question?’ and the person says, ‘I’m very sad, I’m crying’ then how will it know whether that person is lying or not? Now you don’t always know if every human is lying in a relationship but if you have a strong pastoral care relationship with that person and you are seeing something that is different, then you can take the question to a more senior rabi and go through a peer review process... There are many more steps to weed out any lies or inconsistencies… [so the law can be applied correctly]

The role of human judgment

Those at the forefront of AI are grappling with many questions, but one of the most challenging philosophical issues under discussion is whether there is something unique about human judgment that AI cannot replicate.  Human judgment concerns a person’s ability to form an opinion, understand something and make a decision. It's a mental process that involves using emotions, intuition and experiences to interpret ambiguous information and understand context. From the moment we wake up each day, we are constantly exercising this judgement. In many instances these decisions are trivial; in some they can be substantial. Most involve a complex series of trade-offs that are influenced by a person’s unique values, experiences and perceptions.

The way human judgment works can be illustrated through the trolley problem, the famous ethical and philosophical thought experiment. The original scenario goes something like this. There is a runaway trolley (a vehicle like a tram) hurtling down a set of tracks. Somewhere in the distance are five workers doing maintenance that are in its direct path. They cannot see the trolley or hear it coming and when they do, it will be too late for them to get out of the way. Their lives will be lost. But you, as a bystander, notice there is a track lever nearby that can switch the trolley onto a new set of tracks. It is possible for you to use that level to divert the trolley; however, by doing so, you will direct the trolley into the path of a lone track worker, killing him instantly. What should you do? Leave the trolley to take its original course or change it? You are asked to make a judgment call.

Individuals exposed to this problem unconsciously go through a weighing process where they consider the costs and benefits of each course of action based upon their personal values and principles. In this version of the trolley dilemma the ‘right’ outcome seems easy because there is consensus around the issue at stake—the value of human life and our onus to protect it—but in other situations, the ‘right’ solution might not be so clear.

These are the kinds of dilemmas that religious leaders like Rabi Alon face all the time as they provide guidance and pastoral and spiritual care. Most religions have a set of core teachings that must be applied literally and absolutely (yet even here there may be exceptions), while the rest come with caveats and exceptions and contextualisations. This is where being a religious leader becomes challenging—there is often disagreement around these ‘grey’ areas of law or practice. One leader may prioritise certain values, principles or precedents over others. They may give primacy to certain texts or give more weight to reason or intuition or context when reaching a conclusion. Things become even more complex when schools of thought change over time or when the world itself changes. Two leaders may look at the same problem and reach entirely different conclusions and still be correct in terms of applying legal or theological tools correctly. While this diversity in thought brings richness and depth to a religious tradition and allows it to evolve over time, it also makes it very difficult for an algorithm to replicate the process. In some cases, the values, assumptions or preferences a person draws upon when exercising judgement are unconscious—the individual may not even know why they have reached a certain decision. Rabi Alon explains it in this way:

the weighting of sources, the weighting of information, how can a computer understand this? It’s hard enough for us human beings to, who have the nuance of the context to understand, to determine which rabbinic source is prioritized over another…. How can an AI engage with this? How does an AI even ask for the nuance?

Congregants trust their leaders with their spiritual welfare and rely deeply on their guidance. Even if the matter at stake is not one of life and death, from their perspective, eternal consequences may be on the line.

Bypassing the impermissible -Dr Sadeq Ansari 

Dr Sadeq Ansari wouldn’t describe himself as tech savy. He only just managed to navigate the technical aspects of his PhD with generous support from his supervisor, who has since retired. But that research led to his involvement in a project called SirahHub, which has captured the interest of individuals working very much at the forefront of emerging technologies, including AI.

Sadeq is an Islamic studies lecturer and a researcher in Islamic sciences, with a particular interest in the intersection of Islamic tradition and modernity. His background lends itself to this interest. He was born in Afghanistan before living for a time in India and moving to Australia. His formative years of education were all here, but as a young adult he’s spent time in Egypt and Turkey, as well as in the US. His Islam is approached through the lens of growing up in the western world alongside exposure to the traditional. He follows western Muslim scholars but also has a deep respect for and interest in Islam’s rich intellectual tradition.

Sadeq’s PhD led to his secondment to a project that involved gathering biographical data about the life of Prophet Muhammad. Traditionally, this information has only been drawn from certain historical texts. However, there is a wealth of information that can be found in other documents—from legal sources to Qur’anic commentary. The project aimed to expose the genre of Sirah studies to an interdisciplinary approach that would see it incorporate data from a broader range of documents. It took almost 10 years and researchers scrutinising more than 4,000 texts to bring it to a stage of completion where the data could be presented to scholars working in the field.

A year ago, the project caught the attention of a group of individuals working in filmmaking and game design who saw the potential for using that data to teach others about the life of the Prophet. Their background in multimedia gave them an expansive vision—3D projections, animated video series, an entire metaverse where people could not only learn about the Prophet’s life and those of his wives and companions but live it at an experiential level.

To achieve this, they needed to start at the very beginning. A world had to recreated, cities designed, faces reconstructed all from textual descriptions of a time long ago. AI helped in this process. They could feed it information about what people looked like, descriptions of the buildings, information about the landscape, the mountains etc. and AI could put together scenes that would eventually be built into video episodes.

In another project they brought back to life some of the key historical figures from the Prophet’s life using interactive AI avatars. A pilot initiative was the recreation of Amr ibn Hisham, one of Muhammad’s key enemies, a notorious clan leader from Mecca who was renowned for leading opposition against Prophet Muhammad and his teachings. Through a mobile interface users can interact with him, much like interacting with a human person. They can see his facial expressions and mannerism, hear his voice and ask him questions about his life.

According to Sadeq, the initial feedback on these projects has been favourable. A group of scholars who viewed the first animated video episode thought it was a great way to teach the religion. However, there was some concern around the depiction of the Prophet’s first wife Khadija. Some individuals felt uncomfortable about seeing her brought to life, given her special significance in Islamic tradition.

Sadeq says the project is, in some ways, testing the waters of Muslims’ tolerance of depicting significant Muslim figures through visual imagery. Traditionally, representing Prophet Muhammad via any form of image has been forbidden. At times this rule has been extended to other important figures (hence the reticence surrounding Khadija); however, in recent years there has been less resistance to depicting figures like the first four caliphs in things like film. Sadeq feels that AI depictions will only open up further possibilities. AI renditions can bypass some of the objections that have been voiced when using actors and actresses to play Islamic figures because they are not real. It is often argued that when an actor takes on the role of someone from the Prophet’s time, people start to associate that persona with that individual. If that actor (or actress) then goes on to play other roles, particularly those that depict non-Islamic values or behaviours, it can create confusion for viewers. However, when AI avatars are project specific and confined to a particular role, that confusion can be avoided.

The opening up of new possibilities is, Sadeq says, one of the advantages of AI. He feels that often the conversation about artificial intelligence is negative. People are understandably afraid because there are risks. But there are also opportunities too, he says:

I’m always of the opinion that you’ve got to ride the wave that comes. Waves can be dangerous. They create rips and so on. You can get knocked down; you can tumble. But they can be fun as well and you can ride them, and you can get places quicker by riding them instead of fighting them. This is an example of riding them, perhaps…

The contribution of religious leaders to emerging discourses on AI

When I spoke to him some months ago now, the Venerable Thich Phuoc Tan had just returned from a peace conference in Japan, where representatives of different faith traditions had come together to discuss some of the world’s current challenges. Amid dialogue about the war in Gaza and the conflict in Ukraine, AI was an important topic of conversation. There was acknowledgement that artificial intelligences are important tools that can contribute to humanity, but they also come with considerable risk. Delegates at the conference were encouraged to sign the Rome Call for AI Ethics a document endorsed by faith leaders, government representatives and private companies, which calls for AI to be developed in a way that serves humanity. Speakers at the conference emphasised that AI is a fundamentally different technology—it is like nothing that we have experienced before—and it will have ramifications for our perceptions of reality and of ourselves.This means, they emphasised, religious leaders have an important role to play in the global conversation about AI. They have a “unique responsibility” to ensure that the pursuit of AI occurs with “moral clarity and ethical integrity.”

Are faith leaders engaging with this challenge? It is evident that leaders of many faith traditions are actively grappling with the reality and implications of AI. The Rome Call for Ethics has, to date, been endorsed by representatives from eleven world religions, encompassing both Eastern and Abrahamic faiths.These men and women have expressed commitment to seeking “the development of an artificial intelligence that serves every person and humanity as a whole; that respects the dignity of the human person, so that every individual can benefit from the advances of technology; and that does not have as its sole goal greater profit or the gradual replacement of people in the workplace.”

This is a huge aim and there are many forces at play in AI’s development that are working both for and against these objectives. Yet embracing this challenge can start small. It can begin simply by, as Thich Phuoc Tan did, exploring and then mastering AI tools for efficiency purposes. Activities can be written sermons generated, texts identified, and each outcome can be examined to determine its strengths and weaknesses compared to a human effort. With time, the shortcomings and weaknesses of AI will be revealed to all of us and those who are ready and prepared to engage with these broader ethical and philosophical issues will be able contribute to the conversation as to how they should be addressed.

What AI will become is still unknown. At this early point in its emergence, we do not know whether (or to what extent) it will replace human function or whether we are heading towards an all-out SkyNet-type scenario where the world is taken over by machines. But what we do know is that religious leaders and their congregants have always been interested in exploring the unknown, the impossible and the difficult to understand. The mysteries of life, death and the universe have always been their domain—and so we look forward to their contribution to what awaits.