Theological translation debate separates wider access from responsibility for meaning

Two essays published by The Gospel Coalition on October 5 frame a disagreement about artificial intelligence and the translation of historical theological texts. Fred Sanders emphasizes the opportunity to make difficult-to-access works available, while James Eglinton argues that the human work of interpretation cannot be treated as an incidental part of moving words between languages. The essays respond to The Faith Received, a theological-text project described by both writers. This article reports their published positions; it does not independently certify the platform’s performance or recommend a service. Sanders describes potential benefits for readers who face barriers involving language, availability and the form of historical sources. He also identifies errors in material he examined. His favorable assessment is qualified: access to source-page scans can support comparison, but it does not make an AI rendering equivalent to an expert human translation. Eglinton focuses on translation as interpretation. He argues that differences of language and expression require decisions and that those decisions leave a translator’s influence on the resulting text. His concern includes readers who cannot evaluate a translated passage against its source language. The distinction helps explain why the writers’ disagreement is larger than whether a particular output contains a mistake. Sanders considers how imperfect but accessible material can contribute to learning. Eglinton asks what readers may misunderstand about their encounter with an author when the interpretive process becomes less visible. Both essays should be read as arguments by identifiable authors. Their examples do not establish a universal rate of accuracy across translation systems. Nor do two positions from one publication constitute a complete survey of scholarly opinion. For readers, the practical questions arise at different stages. What text is being encountered? Which edition supplied it? Who or what produced the rendering? What knowledge is available to evaluate a consequential phrase? These questions do not require a predetermined verdict about every use of artificial intelligence. The essays also distinguish initial access from authority for a published claim. Finding a readable rendering can open inquiry. Quoting it as evidence of precisely what a historical writer argued requires attention to the source and translation. Sanders places his assessment within a longer effort. He writes that in 2014 he served on the advisory board of an organization committed to retrieving the riches of the Christian tradition for Protestant readers, and that the group repeatedly met the same obstacle: important patristic, medieval and even Protestant works existed only in Latin. The options, as he describes them, were more productive expert translators or more pastors and seminarians able to read Latin for themselves. He judges the gains from both strategies to have been modest, and describes AI translation as having emerged rather suddenly as a third force. His examples are concrete. In one passage, he writes, the AI rendering has Hugo Grotius saying that “we owe God grace” for his blessings, a misreading of gratia, which in that context must mean “thanks.” He treats that error as a warning that other infelicities are likely. What he values is the site’s design, which lets readers click through to a page scan of the original. That feature, he argues, allows comparison with the source, supports citation of a published edition and page number rather than a website, and reminds readers that the screen text is a derivative product. Sanders also states the limit plainly: such transparency “doesn’t suddenly raise an AI translation up to the same level as a human translation by an expert.” For churches and students, the exchange underlines a principle older than the technology. Christians receive the Scriptures and the church’s teachers through the work of faithful translators, and responsibility for meaning remains a human and pastoral task that a tool can assist but not assume. The debate’s value is in making that responsibility visible. Wider access and accountable interpretation need not be reduced to a contest between enthusiasm and refusal. Readers can examine the specific material and the basis on which they understand it, while preserving the difference between an invitation to read and a verified scholarly conclusion.