ARTICLE
21 March 2025

AI In The Religious Sphere: Regulate, Ban, Or Defend?

Ka
Khurana and Khurana

Contributor

K&K is among leading IP and Commercial Law Practices in India with rankings and recommendations from Legal500, IAM, Chambers & Partners, AsiaIP, Acquisition-INTL, Corp-INTL, and Managing IP. K&K represents numerous entities through its 9 offices across India and over 160 professionals for varied IP, Corporate, Commercial, and Media/Entertainment Matters.
With AI now capable of generating religious sermons, interpretations of holy texts, and even prayers, legal and ethical concerns arise. Religious laws in many countries penalize blasphemy or offensive religious speech.
India Technology

With AI now capable of generating religious sermons, interpretations of holy texts, and even prayers, legal and ethical concerns arise. Religious laws in many countries penalize blasphemy or offensive religious speech. If AI generates content that is deemed offensive, who is responsible? The developer, the user, or the AI itself?

Introduction

Imagine that whenever you ask AI to pray, preach, or interpret a holy book, it gives you an instant response. AI-generate religious content is now reality, not science fiction. Chatbots powered by AI give spiritual advice to algorithms that script sermons and hymns. As AI starts to take over religious discourses, it also brings up several moral and legal issues before it, such as who owns the content? Can AI commit blasphemy? And should faith produced by it be regarded as digital fiction or as divine?

Religious texts, sermons, prayers, or theological interpretations made or modified or delivered by AI is known as AI-generated religious content.

GitaGPT is the most famous of these AI-based chatbots. It offers answers to queries by users in a tone that echoes Lord Krishna.1 Built on the OpenAI technology of GPT-3, GitaGPT offers scriptural advice for moral dilemmas, personal struggles, and philosophical inquiries. However, the rising popularity of GitaGPT and similar tools has led many to question whether AI-generated religious interpretation can ever be as authentic or accurate.

While freedom of religion is a cardinal right under the international human rights law, still many countries employ restrictions through their blasphemy laws to prevent giving religious offense. Key international legal provisions include :

Universal Declaration of Human Rights (UDHR), 1948 – Article 18 guarantees the right to freedom of thought, conscience, and religion, while Article 19 protects freedom of expression. However, these rights are often restricted in cases where religious sentiments are affected.2

International Covenant on Civil and Political Rights (ICCPR), 1966 – Article 18 protects religious freedom, while Article 19 allows restrictions on free speech to protect public order and morality.3

Several legal provisions regulate religious speech and protect religious sentiments:

Article 25 guarantees the freedom of religion, subject to public order, morality, and health.4

Section 295-298 of the Bharatiya Nyaya Sanhita (BNS) – These sections criminalize actions that insult or harm religious sentiments.5

Section 302: Uttering words to deliberately wound religious feelings.6

Section 69A Information Technology (IT) Act, 2000 allows the government to block online content that threatens religious harmony.7

Section 353 BNS: Punishes statements that incite communal disharmony.8

Recent Cases & AI-Generated Religious Content –

Cases like Ramji Lal Modi v. State of U.P. (1957)9 upheld Section 295A erstwhile IPC, affirming restrictions on religious speech.

With the rise of AI-generated religious content, concerns arise about whether AI-created texts or imagery could violate Section 299 BNS or be deemed "deliberate and malicious."

This indicates that although India guarantees religious freedom, its legal system imposes severe restrictions on religious speech to maintain harmony.

Liability in Religious Speech: Who is accountable—developers, users, or AI?

Because of the lack of conscious intent of AI, its application in relation to traditional legal frameworks is highly challenging. There is a strong debate on whose liability should fall: the developers who created it, the users who gave it prompts, or the AI itself?.

In turn, developers must make AI models. They need to train AI models on a certain dataset with embedding safeguards in them to not post harmful content. They outline the ethical parameters of the system and might be held partially culpable for not installing adequate content moderation. However, since machine learning has its unpredictable nature, it is unreasonable to hold developers accountable for every unexpected consequence

Users, directly impact the AI-generated religious speech by using their prompts. If a user intentionally asks an AI to produce controversial or blasphemous religious content, then they may be liable under Indian laws relating to hate speech and religious harmony such as Sections 153A and 505 of the IPC. Courts in India have convicted people for publishing offending religious content on the internet, but that raises the question of whether the same rules could be applied to AI-generated material

The most challenging question is whether AI itself would be liable. At present, no legal system in the world considers AI a separate legal entity and therefore can neither punish it nor sue it.

But debates are emerging on AI personhood also in intellectual property law, and some legal scholars even argue that criminal liability may eventually extend to AI systems.

It can be said that the current law of India has not addressed this issue of AI-generated religious speech, and so courts have been interpreting laws developed for human actors. As a new Digital India Act is currently being worked out, it is essential to outline the liability of AI platforms in religious and blasphemy-related cases. Regulatory models could draw inspiration from the EU AI Act (2023), which imposes strict obligations on AI developers and platforms regarding harmful content.

Freedom of Expression vs. Religious Sensitivities: Can AI-generated religious content be protected under free speech laws?

The question that should such contents still be subjected to the same legal constraints of speaking in a human voice, with AI itself devoid of any conscious intention to create slanderous speak or subsume any religious viewpoint?

Some in this paradigm favour complete and unrestricted social protections of free speech for what AI-generated religious expression happens to hold. They argue that AI-generated religious content stands in the same sphere of free expression as all other digital content and would not need additional regulation at all. However, this position could brush against unsubtlety, which could many times incite polarizing sentiments among faithful faith communities and lead to serious social and legal ramifications.

On the other end of the spectrum, the regulated expression approach opens the gates-of- holy blasphemy- to AI-generated religious content and in so doing ensures, on the one hand, that it does not incite religious tensions and offend believers. This model prioritizes social cohesion but raises concerns about excessive censorship and restricting technological development.

Another possible solution would be a hybrid approach. Here, it would propose the idea of giving free speech rights to AI-generated religious content conditionally. It may still offer an outlet for AI content dealt with in a discursive form by leaning over backwards to accommodate religious sensitivities and possibly prevent conflicts.

"Deus in Machina"

The case of "Deus in Machina" at St. Peter's Chapel in Lucerne, Switzerland with a digital Jesus avatar permitted the users to seek guidance through AI-generated responses derived from the Scriptures.10 Some viewed it as insightful and spiritually uplifting, while others called it blasphemous, protesting that no AI should impersonate holy beings.

The question here arises that, "should AI present itself as any religious figure?" "Would this be construed as violating religious sentiments?"

Ethical AI regulations to prevent offensive religious speech.

Legal accountability and liability is crucial, but the design of AI in generating responses and decisions can cause religious conflicts. AI systems that engage with religious matters must take into account cultural and religious sentiments. Algorithmic fixes may additionally reduce the risk of generating blasphemous or otherwise offensive religious content, allowing responses to remain within ethical lines that are generally deemed appropriate.

Another equally important component of responsible AI development is transparency. AI platforms should also be required to cite their sources of training data, particularly in relation to religious texts and interpretations. Further, disclaimers ought to be included lest the user gets the idea that what an AI generates in the name of religion is authoritative doctrine rather than a computational interpretation.

Conclusion

There can be seen increasing use of AI in religious interferes. Tools such as GitaGPT give rise to questions of authenticity, personal intent, and legal liability. The working of international frameworks such as the UDHR (1948 and the ICCPR (1966) are to protect religious liberty but allowing of certain limitations. One of the big legal challenges is liability of whom the onus of responsibility is on AI-generated religious content? Considering the AI is imperfect in intent, does accountability fall on developers, users, or platforms? It will take a balanced legal stand to avoid over-censorship as well as freedom without limits that may cause harm.

The approaches taken by a countries have varied. Indian, Pakistani, Saudi, or European cases show that some are very strict on defamation laws regarding religion, while others support free speech exactly.

A hybrid model with a balanced legal framework capable of constraining AI without stifling innovation is, to an extent, the only practical answer.

Footnotes

1. sneha, 'Google Software Engineer Creates Gita GPT, an AI Chatbot Inspired by Bhagavad Gita' (9 February 2023) (https://www.indiatoday.in/technology/news/story/google-software-engineer-creates-gita-gpt-an-ai-chatbot-inspired-by-bhagavad-gita-2332576-2023-02-09) accessed 7 February 2025

2. 'Universal Declaration of Human Rights' (United Nations) (https://www.un.org/en/about-us/universal-declaration-of-human-rights) accessed 7 February 2025

3. (International Covenant on Civil and Political Rights | Ohchr) (https://www.ohchr.org/en/instruments-mechanisms/instruments/international-covenant-civil-and-political-rights) accessed 7 February 2025

4. The Constitution of India, 1950, art 25

5. Bharatiya Nyaya Sanhita, 2023

6. Bharatiya Nyaya Sanhita, 2023, s 302

7. Information Technology (IT) Act, 2000, s69A

8. Bharatiya Nyaya Sanhita, 2023, s 353

9. 1957 SCC 334

10. Keaten J, '"ai Jesus" Avatar Tests Man's Faith in Machines and the Divine' (AP News, 28 November 2024) (https://apnews.com/article/artificial-intelligence-chatbot-jesus-lucerne-catholic-66268027fbcf4b48972d1d62541f0b16) accessed 7 February 2025

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

Mondaq uses cookies on this website. By using our website you agree to our use of cookies as set out in our Privacy Policy.

Learn More