Home Scripture Forge

The Impact of AI on Bible Translation: Opportunities and Challenges

AI image: Google Gemini

By Taeho Jang

Forty years ago, I began studying artificial intelligence. Machine translation, a specialized field within AI, sparked my initial interest in Bible translation. At that time, I was fortunate to study early AI tools designed specifically for Bible translation work. My career eventually shifted toward becoming a translator and linguist, allowing me to experience the complete process of traditional Bible translation—from the phonological analysis of the target language to the completion of the New Testament. This journey provided me with valuable insider perspectives on Bible translation while reinforcing my ongoing interest in applying computational technology to the field.

The rule-based translation systems of that era presented significant limitations. They required human experts to explicitly define linguistic rules, but I quickly discovered that the inherent complexity of language made it virtually impossible to formulate comprehensive rule sets that could adequately capture all linguistic phenomena.

A watershed moment

When AlphaGo defeated a human champion at Go in 2016, I sensed a profound technological breakthrough. This moment signaled a key advancement in the Fourth Industrial Revolution—the development of sophisticated language models that accurately simulate human language processing. These innovations are fundamentally transforming language-related fields, particularly translation, and Bible translation is no exception. This AI revolution reignited my decades-old vision of leveraging AI for Bible translation.

The Third Industrial Revolution's information age created standardized biblical text databases that now serve as crucial foundations for implementing data-driven AI translation technologies. However, many languages still awaiting translation have limited linguistic data, creating challenges for AI application. We addressed this obstacle by transferring translation intelligence from language models developed for resource-rich languages. This approach resulted in the Scripture Forge AI drafting product, which was released in 2024 and is currently implemented or in preparation in nearly 500 projects around the world.

Feedback from my AI translation workshop participants has been overwhelmingly positive. Teams that previously completed New Testament translations and subsequently applied AI drafting to Old Testament books report “amazing” quality results. For translation environments where sustaining human translator engagement is challenging, AI drafting provides valuable, continuous assistance. However, results vary. Projects with limited or poor-quality training data may experience suboptimal outcomes, resulting in the need for expert consultation.

AI’s advantages ... and risks

AI translation's primary advantage is its ability to produce naturally flowing language, which often surprises mother tongue translators (MTTs) experiencing AI drafting. However, this naturalness can mask risks when AI prioritizes fluency over fidelity, occasionally using expressions that diverge from the source text’s meaning. These localized errors hidden within otherwise fluent text represent potential risks in AI translation.

Critical evaluation skills are essential for identifying these subtle errors and effectively utilizing AI drafts. Translation teams that adopt AI without sufficient critical assessment capability may overlook nuanced inaccuracies, potentially leading to unintended consequences. Comprehensive understanding and training are therefore crucial safeguards. These same principles apply broadly when integrating AI technologies into church contexts.

The fundamental change AI brings to society can be characterized as resource decentralization. This shift is particularly significant for Bible translation—a long-term ministry dependent on MTTs’ capabilities and sustained availability. Languages still awaiting translation typically exist in resource-constrained environments. AI technology offers a revolutionary solution by dramatically increasing translation resources for these underserved language communities, providing renewed hope for completing translations for all remaining languages.

Intelligence forms the essence of AI and drives the Fourth Industrial Revolution. During the Third Industrial Revolution, humans retained decision-making authority over information processing. However, as the current revolution progresses, AI increasingly assumes decision-making functions. The Bible translation process involves countless decisions across multiple domains—from MTTs' word selection to consultant checking. AI's role in this decision-making landscape carries profound implications.

AI transforms human roles

When AI handles simple, repetitive drafting decisions, translation efficiency increases substantially. Nevertheless, I firmly believe that humans should retain final authority over high-level decisions, particularly those involving exegesis and theological nuance. As AI capabilities rapidly advance, we must resist the misconception that AI can independently complete Bible translation. The essential characteristic of Bible translation is its spiritual dimension—a process guided by the Holy Spirit through prayer. We must vigilantly preserve this spiritual aspect amid technological advancement.

The current Scripture Forge AI drafting system aims to produce only initial drafts (draft 0), with subsequent processes following established translation protocols involving human oversight. Some express concern that AI drafting might deprive MTTs of skill-development opportunities traditionally gained through manual drafting. However, this technological shift actually presents MTTs with a new opportunity for professional advancement. AI fundamentally transforms the translation paradigm—elevating MTTs' roles from drafting to critical evaluation and refinement of AI-generated drafts.

Importantly, high-quality manual translation experience remains essential before implementing AI tools. This practice provides translators with the necessary expertise to effectively evaluate AI-generated content. The critical thinking skills required to identify subtle errors in AI drafts build upon existing translation competencies. Another reason why this manual translation is necessary is because a certain amount of training translation data is required to utilize AI.

The opportunity to leverage AI capabilities is now available to teams that have already completed substantial translation work. Translation teams interested in exploring AI implementation can sign up for the Scripture Forge program, click the Generate Draft button and sign up for drafting to receive personalized guidance on integration options.

By carefully balancing technological innovation with human expertise and spiritual sensitivity, Bible translation can harness AI's powerful capabilities while maintaining translation fidelity.

•••

Dr. Taeho Jang of Global Bible Translators (GBT) in South Korea is a translation consultant actively engaged in AI-assisted Bible translation. He helps many translation teams through AI translation workshops and seminars. Contact him at taehojan@gmail.com

News

View all articles

Special Report - AI, Bible Translation & the Church

‘We’ve come very far, very fast’

A tech observer outlines what AI will mean soon for workplaces and ministry

Tech pioneer: Christians ‘have to show up’ for AI

Silicon Valley pioneer Pat Gelsinger was CEO of Intel Corporation until December 2024. Quickly realising his career in technology was not finished, he joined the faith/tech platform Gloo in early 2025 as the executive chair and head of technology. He is also a general partner at the venture capital firm Playground Global. Gelsinger was instrumental in the development of cloud computing, Wi-Fi, USB and many other everyday technologies. He estimates his work has touched 60 to 70 percent of humanity. Here are highlights of his keynote talk at the 2025 Missional AI Summit. You can watch his entire talk here. Pat Gelsinger (left) is interviewed onstage by Steele Billings. Both are with Gloo. Watch the full interview here. Is technology good or bad? Technology is neither good nor bad. It’s neutral. It can be used for good. It can be used for bad. … If you think back to the Roman roads, why did Christ come when he came? I’ll argue the Pax Romana and the Roman roads. … The greatest technology of the day was the Roman road system. It was used so the Word could go out. Historical example I will argue Martin Luther was the most significant figure of the last thousand years. And what did he do? He used the greatest piece of technology available at the day, the Gutenberg printing press. He created Bibles. … He broke, essentially, the monopoly on the Bible translations …. He ushered in education. He created the systems that led to the Renaissance. That’s a little punk monk who only wanted to get an audience with the pope because he thought he had a few theological errors. I’ll argue (Luther was) the most significant figure of the last thousand years, using technology to improve the lives of every human that he touched at the time. How today compares to the dawn of the internet AI is more important. AI will be more significant. AI will be more dramatic. … This is now incredibly useful, and we’re going to see AI become just like the internet, where every single interaction will be infused with AI capabilities. In the 75-year-or-so history of computing, we humans have been adapting to the computer. … With AI, computers adapt to us. We talk to them. They hear us. They see us for the first time. And now they are becoming a user interface that fits with humanity. And for this and so many other reasons that every technology has been building on the prior technology, AI will unquestionably be the biggest of these waves, more impactful even than the internet was. On the need for AI development to be open-source It is so critical because we’re embedding knowledge, embedding values, embedding understanding into those underlying models, large language models and every aspect that happens. It must be open, and this is part of what I think is critical about us being together here today. We need to be creating trusted, open, useful AI that we can build humanity on. On the need for Christians to help build AI systems We have to show up as the faith community to be influencing those outcomes, because remember what happened in the social media. We didn’t show up, and look at what we got. So are we going to miss this opportunity for something that’s far more important than social networking with AI? Where it truly in the models embeds every aspect of human history and values into it? We have to show up, team. What we do with large language models is far more important because truly we are choosing how we embody knowledge of all time into those underlying models. They need to be open. They need to be trusted. What Christians must bring to the process If we’re going to show up to influence AI broadly, we have to show up with good engineering, good data, good understanding, good frameworks. How do you measure things like ‘Is that leading to better character? Is that leading to better relationships? Is that creating better vocational outcomes? Is that a valid view of a spiritual perspective?’ We need good underlying data associated with each one of these. And for that we’re actively involved. We’re driving to create that underlying data set. Because we need to show up with good data if we’re going to influence how AI is created. How should this work? For the AI systems we need to create good benchmarks. If I ask about God, does it give me a good answer or not? If I ask about relationships with my children, does it give me good answers? We need to create the corpus of data to give good answers to those questions. And, armed with that good data, we need to show up to influence the total landscape of AI. We want to benchmark OpenAI. We’re going to benchmark Gemini. We’re going to benchmark Claude. We’re going to benchmark Copilot. This is what we’re going to do at Gloo, but we want to be part of a broader community in that discussion so that we’re influential in creating flourishing AI. Technology is a force for good. AI that truly embeds the values that we care about, that we want to honour, that we want to be representing into the future and benchmarking across all of them. Oh his role with Gloo We are going to change the landscape of the faith community and its role in shaping this most critical technology, AI, for faith and flourishing. That’s what we’re going to do at Gloo and we need all of your help and partnership to do so because if we don’t hang together, we’re not going to influence the outcome, right? ‘Here am I, Lord’ I don’t think I’m done. … You and I both need to come to the same position like Isaiah did. Here am I, Lord. Send me. Send me. Send us. That we can be shaping technology as a force for good. That we could grab this moment in time. This is the greatest time to live in human history. We’re going to solve diseases. We’re going to improve lives. We’re going to educate every person in poverty. We are going to solve climate issues. We are going to be using these technologies to improve the lives of every human on the planet. We are going to shape technology as a force for good. Here am I, Lord. Send me. ••• Story: Jim Killam, Wycliffe Global Alliance Translated with ChatGPT. How was the translation accuracy? Let us know at info@wycliffe.net. Alliance organisations are welcome to download and use images from this series.