Data Safety for LLMs in Bible Translation

In this episode of the Bible Translation Innovation Podcast, Joel Mathew, Chris “Klappy” Klapp, and Isabella Scarinzi explore the growing security and privacy challenges surrounding the use of AI and large language models in Bible translation, particularly in sensitive contexts. The discussion examines the distinction between privacy and security, the risks of sharing sensitive information with AI systems, and how data collection, model ownership, and government influence can affect trust in AI tools. Joel and Klappy also discuss emerging threats such as AI-enabled phishing, deepfakes, voice cloning, and highly connected AI systems that may unintentionally expose sensitive data. The episode also offers with practical recommendations for organizations, including establishing clear AI policies, evaluating the provenance and ownership of AI models, limiting AI access to sensitive information, and balancing the tremendous potential of AI-assisted Bible translation with the responsibility to protect translators, teams, and communities around the world.Have questions? Send us an email at [email protected] and we'll answer them on the show!Subscribe to hear more conversations, updates, and experiments on new methodologies and technologies advancing Scripture accessibility worldwide.