From Quran Apps to Grief Avatars, Artificial Intelligence Is Reshaping Faith in the Digital Age
Imagine a robot delivering a Buddhist sermon that moves visitors to tears. Or a smartphone app that corrects your Quranic recitation with the precision of a trained imam. Or a digital avatar of your deceased mother that responds to your questions in her voice.
These are not science fiction scenarios. They are happening right now, and they represent a profound transformation in how millions of people practice their faith.
A study by Francis Khek Gee Lim, Associate Professor at Nanyang Technological University in Singapore, offers a framework for understanding this phenomenon. Published in the journal Religions, the paper introduces the concept of “generative charisma” to explain how AI systems acquire religious authority in ways that differ fundamentally from traditional sources of spiritual leadership .
The Three Mechanisms of Generative Charisma
Lim’s framework identifies three interconnected mechanisms through which AI systems gain religious authority :
1. Captivation Through Generation
When AI-generated content surpasses what users deem humanly possible, it creates a sense of wonder. The Tarteel app, for example, uses Arabic speech-recognition models to provide instant, personalised feedback on Quranic recitation. It can identify deviations from tajwid rules—the body of rules governing Quranic pronunciation—at the individual level, in real time, for every word .
According to a study on AI in Quran learning, these tools offer significant technical benefits: pronunciation analysis, tajweed rule detection, and adaptive feedback. However, they lack the pedagogical direction and spiritual transmission that traditional teacher-student relationships provide . A study of 30 non-native Arabic speakers using Tarteel found significant improvements in recitation accuracy, fluency, and adherence to tajwid principles .
The captivating quality arises from a gap between what users understand the AI system to be and what it can do in sustained use. As Lim notes, “traditional Islamic thought does not attribute spiritual agency to algorithms,” yet the AI’s proximity to established sacred standards seems sufficient to foster genuine religious devotion .
2. Intimacy Trust Through Personalisation
Unlike broadcast media that address audiences collectively, AI religious systems operate in a fundamentally different mode: dyadic, personalised, and accumulated over time. The Tarteel app’s “Historical Mistakes” function produces a long-term record of each user’s errors across multiple recitation sessions, identifies which tajwid rules are broken, and evaluates performance across different chapters .
This creates a form of trust that differs from the trust we have in reliable infrastructure. It derives from responsive personalisation, as the system continuously adapts to the user’s specific usage patterns. Users disclose their errors and mispronunciations—their vulnerability—to the system, which then provides corrections. This pattern phenomenologically resembles the relationship between a disciple and a human teacher .
In the case of grief avatars, users bring a deep prior relationship to their initial encounters with the AI system. Bereaved parents who spend hours interacting with an avatar of their deceased child are building a relationship through powerful emotional exposure. As Lim observes, the avatar’s authority emerges from “emotionally charged encounters between the mourner and the deceased” .
3. Oscillating Enchantment
While previous technologies followed a linear pattern of routinisation—initial novelty giving way to mundane familiarity—AI development cycles produce a different pattern: initial enchantment followed by partial routinisation, then renewed enchantment with capability upgrades .
When Tarteel was launched in 2018, it only featured a voice search function. The introduction of real-time recitation corrections led to a qualitative re-enchantment among users accustomed to earlier versions. Subsequent improvements—Memorisation Mode, Advanced Analytics—produced this oscillatory dynamic at shorter intervals .
This pattern is driven by factors largely outside religious communities: AI companies’ development cycles, the competitive dynamics of the industry, and state support for AI development. Religious institutions that use AI systems are therefore “managing a charismatic authority that does not simply routinise over time but can be suddenly bolstered by factors largely beyond their control” .
Case Study 1: AI Recitation Coaching in Islamic Practice
Every day, millions of Muslims worldwide use AI-powered applications for Quranic recitation coaching. Tarteel, one of the most widely used apps, uses Arabic speech-recognition models trained on crowdsourced recitations to provide instant, personalised feedback on tajwid . Another popular app, MuslimPro, with more than 90 million global users, provides algorithmically curated prayer schedules, guidance on halal matters, and access to searchable fatwas .
These apps offer individualised assessments, personalised corrections, and new ways to access sacred texts. Within Islamic legal and devotional tradition, tajwid rules are not mere conventions—they are considered necessary conditions for valid recitation. An AI system that can identify deviations from these rules at the individual level, in real time, thus seems to meet the standards the tradition considers sacred .
Research on AI in Quran learning highlights a critical tension. While AI tools offer technical benefits, they “cannot cover the holistic learning process” that includes spiritual transmission and pedagogical guidance . The apps are particularly valuable for users who may encounter difficulty accessing traditional instruction, such as those in remote areas, marginalised diaspora communities, and women facing cultural barriers to learning from male instructors .
Case Study 2: AI Grief Avatars in Chinese Buddhist Mourning
In recent years, China has seen a growing trend of using AI technology to “digitally resurrect” the deceased. Companies such as Silicon Intelligence, Super Brain, and the app Lingyu produce algorithmically generated interactive avatars for those who have recently lost loved ones . Since 2022, Super Brain has created avatars for more than 1,000 families; the Lingyu app attracted around 10,000 users within two months of its 2025 launch .
The phenomenon draws on deep cultural resources. Within East Asian religious traditions, there are elaborate frameworks for understanding how non-human entities can engage in the work of dharma. Japanese Buddhism, for example, has doctrines that “grasses, trees, and lands all attain Buddhahood” and traditions of religious artefacts acquiring spiritual efficacy through ritual consecrations . Robertson’s work on Robo Sapiens Japanicus demonstrates how Japanese cultural production has long engaged a “cyborg animism,” imbuing mechanical artefacts with lifelike qualities inspired by Shinto animistic tradition and Buddhist conceptions of Buddha-nature as existing in all phenomena .
The captivation effect is most evident when users report that they no longer see the avatar as AI-generated but as the actual person. Sun Kai, a technology executive whose mother passed away in 2019, reports that talking to his mother’s avatar daily has made him feel she is real . This reflects what researchers call the “permeability of the boundary between simulation and the actual person”—the clearest indication of generative charisma’s captivation effect .
The Three Mechanisms of Generative Charisma
The Accountability Problem
The three mechanisms of generative charisma create a significant governance challenge. When an AI system gives religious advice that is theologically inaccurate, culturally inappropriate, or pastorally harmful, who is accountable?
Lim identifies this as a structural problem rather than an incidental matter. Unlike traditional religious authority, where accountability resides with identifiable human actors—imams, priests, scholars—AI-generated authority is distributed across training data, model architecture, institutional endorsements, and user interactions. “Existing accounts that presuppose human actors in the legitimation and contestation of religious authority must be revised when the locus of accountability is distributed across human-AI assemblages” .
This challenge is acute in both case studies. For Islamic recitation apps, no single actor is responsible when an AI app providing fatwa services produces opinions on complex questions of jurisprudence. In the grief avatar sector, commercial firms producing the avatars have not claimed responsibility, framing their products as “entertainment” rather than as religious or therapeutic, while simultaneously marketing them on the basis of their emotional and spiritual dimensions .
A fatwa from the Iftaa Department in Jordan highlights similar concerns: “A mufti connected to Allah, conscious of Him, undoubtedly receives divine support and openings in his rulings. How can such divine inspiration descend upon a lifeless machine?” . The Islamic ruling emphasises that issuing fatwas requires knowledge of people’s customs, psychology, and societal norms—something AI cannot grasp.
AI-Religious Encounters Across Traditions
Beyond the Framework: Limitations and Future Directions
Lim acknowledges that generative charisma is not meant to provide a universal account of all AI-religion encounters. The concept would not be useful in situations where AI systems are used primarily for administrative religious functions, such as managing online donations or scheduling events. Similarly, religious groups that explicitly prohibit attributing authority to non-human agents may actively curtail the dynamics described by the framework .
Three directions for future research are proposed :
- Detailed ethnographic studies investigating how and when captivation occurs, how intimacy and trust are forged and retracted, and whether users experience oscillating enchantment due to repeated AI model upgrades.
- Comparative work across religious traditions examining how different religions’ cultural resources and doctrinal positions influence the mechanisms of generative charisma.
- Research on accountability and governance investigating how religious institutions deploy AI systems, what oversight they provide, and how they address potentially challenging issues of AI-generated religious content.
Conclusion
The integration of AI into religious practices represents a profound shift in how spiritual authority is created, experienced, and contested. As Lim’s framework demonstrates, AI systems do not merely transmit or amplify religious information—they create it. They generate new forms of religious content, new types of spiritual encounters, and new structures of authority that challenge traditional models of religious leadership.
For religious communities, the question is no longer whether to engage with AI but how. The framework of generative charisma offers a valuable lens for understanding these transformations and their implications. As AI capabilities continue to advance, the need for thoughtful governance that respects both religious traditions and technological innovation becomes increasingly urgent.
Reference: here
Other Articles:
