Faith in the Age of AI: Lessons from IslamiCity

In a world where artificial intelligence is shaping everything from our finances to our friendships, how can ancient wisdom guide modern technology? That question sits at the heart of the AI and Faith podcast, a platform that brings together technologists, professionals, and religious leaders to discuss the ethical dimensions of AI through the lens of centuries-old moral traditions.

In this episode, host Gilad Berenstein sat down with Mohammed Abdul Aleem, co-founder of IslamiCity, and Zayd Simjee, AI advisor at IslamiCity and co-founder of Guardrails AI, to discuss their mission: using AI to make authentic Islamic knowledge accessible to everyone - while keeping humility and responsibility at the core.

From Bulletin Boards to Digital Wisdom Houses

IslamiCity's story began in 1995, long before AI or smartphones. Back then, Aleem and his co-founders ran an Islamic bulletin board system - accessed via dial-up modem - to share religious resources. When the web browser emerged in 1994, they leapt into the new medium, registering some of the earliest domain names (Islam.org and Allah.org) in Islamic education.

Their vision: a "digital House of Wisdom" to connect people to their faith. Over the years, the platform has grown into one of the most trusted online sources on Islam, providing searchable Qur'an and Hadith databases, linguistic resources, and scholarly articles for both Muslims and non-Muslims.

Enter ChatILM: An AI for Islamic Knowledge

Simjee volunteers for the organization to advise on AI initiatives, the first of which was ChatILM - "ILM" being the Arabic word for knowledge. The goal was simple but ambitious: provide clear, trustworthy answers to everyday Islamic questions in plain English, backed by reliable sources.

"We're not replacing scholars," Simjee stresses. "We're enhancing the baseline knowledge so people can ask better, more relevant questions."

The chatbot draws only from vetted sources such as the Qur'an, Hadith collections, and IslamiCity's own curated library. This helps avoid the common pitfalls of general AI models, which may produce correct-sounding but poorly sourced or sect-specific answers without context.

Bias, Language, and Inclusivity

One of the biggest challenges? AI bias.

While large language models like ChatGPT can handle many general queries well, they often fall short in religious contexts. In Islam, sourcing is paramount - and without proper citation, even correct answers can be misleading. Add in translation layers between Arabic and English, and the risk of error compounds.

The team also faces the complexity of intra-faith diversity. Islam encompasses multiple sects and sub-sects, each with its own jurisprudential nuances. ChatILM aims to be inclusive by referencing appropriate sources depending on the user's background, while clearly distinguishing facts from interpretation.

When AI Should Step Aside

Both guests emphasized the importance of human escalation for certain topics - especially personal, sensitive, or potentially harmful questions.

"If someone asks about divorce, health, or a crisis situation," Simjee explains, "sometimes the right move is to direct them to a scholar, counselor, or even emergency services. AI has limits."

For the majority of everyday questions, though, the system can provide accurate, non-controversial guidance, freeing human experts to focus on more complex issues.

Beyond Chatbots: New Horizons for AI in Faith

Looking ahead, IslamiCity sees potential for AI far beyond Q&A. Ideas include:

• Co-pilot tools for religious scholars and educators, streamlining research and source retrieval.
• Faith-aligned financial guidance, helping Muslims navigate investing, zakat (almsgiving), and avoiding prohibited transactions.
• Persona models trained on the works of historical scholars like Al-Ghazali or Rumi, offering users an interactive way to explore classical thought - with the clear caveat that these are simulations, not real authorities.

Bridging the Tech Gap in Religious Communities

Both Aleem and Simjee acknowledge that religious organizations often lag behind businesses in adopting technology, partly due to limited budgets. Their advice: seek non-profit grants from major cloud providers, experiment with smaller self-hosted models, and engage directly with emerging tools - before someone else builds them without community input.

"If we don't understand these tools," Aleem warns, "we're giving that authority to someone else."

Responsibility at the Core

For Aleem, responsibility flows from faith itself:

"If we believe in God, we must use knowledge ethically. That means making sure these technologies benefit people - and holding corporations accountable to the same."

Simjee adds that transparency in sourcing and a focus on bringing people closer to God should guide every technical decision.

The Human Future of AI

In the end, both guests see AI not as a replacement for religious leaders but as a force multiplier - one that can elevate conversations, deepen understanding, and connect people across boundaries of geography and sect.

"We're only going to make scholars' jobs more difficult," Aleem says, "because people will be asking more challenging questions."

That, they agree, is the kind of challenge worth embracing.