Guidelines for New Testament Draft Generation
Overview
Scripture Forge has successfully deployed AI-assisted draft generation to accelerate preparation of first drafts for Old Testament translation projects. This established capability has now been extended to benefit New Testament translation projects. This innovation can significantly reduce the time required for a translator to prepare a high quality first draft of their New Testament book.
Effective use of Scripture Forge draft generation for New Testament projects depends on several key considerations:
- At least one completed Gospel is required to get started.
- Some book sequences perform much better than other sequences for draft generation.
- Understanding Scripture Forge’s new incremental draft generation capability and estimated draft quality is important to maximize benefits from New Testament draft generation for the Gospels and Acts.
This document provides essential guidance for using Scripture Forge draft generation in a New Testament translation project. The guidance offers insight into the likely draft quality for different New Testament books at various stages of a project, and explains how to get the most benefit from incremental draft generation and draft quality estimates.
Audience
This guidance document is designed for those involved in the planning, execution, and quality control of New Testament projects using Scripture Forge draft generation.
Primary Audience
- Project Planners: To understand the effect of book sequencing on draft quality, and to anticipate the expected changes in draft quality when planning timelines and resource allocation.
- Team Leads: To make informed, practical decisions when selecting the next book for translation to maximize the benefits of AI drafting.
Secondary Audiences
- Translators and Reviewers: To understand draft quality estimates and how book sequencing and incremental draft generation impact the quality of the drafts they will be working with.
- Translation Consultants: To advise teams on the best implementation strategies for using draft generation.
- Trainers: To understand the draft generation methodology for New Testament projects in order to effectively train translation teams.
Background
Scripture Forge draft generation uses Artificial Intelligence to learn from your previously translated Scripture and to provide rough drafts of new books for your Paratext project. These rough drafts can be edited or consulted by the translation team to create a first draft, saving valuable time and effort while improving quality.
Draft generation is enabled by three important types of data - your reference project, your Paratext project, and your drafting source.
- Reference project
The DBL resource or Paratext project that your translation team primarily consults as they translate (i.e., your “model text”). - Paratext project
Completed books from your Paratext project. Especially for New Testament projects, the specific completed books in your project influence generated draft quality. - Drafting source
The DBL resource or Paratext project that will be used for generating drafts. This is typically the same as the Reference Project.