The Gospel Coalition tightens its rules on AI-written articles

The Gospel Coalition has finalized new guidelines that sharply restrict artificial intelligence in the writing it publishes. The Christian media group says all written content on TGC should be the original work of its author, and that if a submission is suspected or confirmed to have been generated wholly or partly by AI, it will not be published. The policy, set out in an AI guidelines document for TGC submissions, says the group holds “the human element of the writing process in high regard, with a low view of AI.” It rejects not only AI-generated drafts but also AI-edited writing, saying editors are not looking for prose that reads like machine output or that has been stripped of a writer’s voice. TGC said the new rules are meant to protect readers’ trust in the integrity of its content and to honor writers by making clear that it publishes only human-crafted work. The guidelines encourage contributors who want editing help to ask another person, such as a well-read friend, a fellow writer or an English-teacher church member, rather than an automated tool. The policy still allows some limited AI use. Writers may use AI for narrow research in the way they might use a basic search engine, but they must verify the results because the tool is “not always reliable.” AI is also barred from the pre-writing stage, including clarifying a thesis or outlining ideas, and it may be used after drafting only to check spelling, not to carry out a full edit. TGC also warned against programs such as Grammarly when they go beyond simple error-checking and start changing language. The editorial line is stricter than some other uses of AI elsewhere in the organization. In an earlier article about the new stance, Collin Hansen said TGC had spent months working with various AI experts and ethicists on guidance for churches, schools and parents, and that reactions to the policy ranged from support to criticism that it was too restrictive. He said TGC was “not anti-tech” but “pro-human,” and would not surrender its mind to machines. The coalition’s own writing on the issue has cast the policy as a missional question as well as an editorial one. In its announcement, TGC argued that as AI-generated text floods the internet, much of what remains will be low-value and that human readers will care most about “distinctly human words” protected from AI’s effects. The group’s position is not entirely new. TGC India’s submission guidelines, published in 2024, already required contributors to declare any generative AI use and to state that no such tool was used for structuring, first-draft composition or substantial style editing. Those rules also said TGC India avoids AI-generated compositions for main content or style edits. Under the new TGC guidelines, undisclosed AI use can have consequences beyond a single submission. If a piece was published with undisclosed AI-generated elements, the editorial director and editor in chief will decide how to rectify the matter and may remove it from the website. If a submission involved undisclosed AI, the editorial director, in consultation with the editor in chief, will decide whether the contributor may keep submitting.