A year ago, machine translation clauses in publishing contracts were a niche concern for technical publishers. This autumn, they are standard battleground. The dividing line running through the translation market is no longer whether to use AI-assisted translation, but under whose control, on what terms, and with what disclosure to authors.
The fault line is clearest in volume genres. Romance, thriller and saga lists with aggressive schedules have quietly adopted MT-plus-post-editing pipelines, arguing that the economics make professional translation viable for midlist titles that would otherwise not be translated at all. Quality controls vary widely, and the better operators now employ dual review — one linguist for accuracy, a second for voice.
Literary publishers have moved in the opposite direction, not by banning the tools but by legislating transparency. New clauses circulating among agents require explicit notice when any machine-generated text enters a translation workflow, give authors approval rights over MT use, and separate MT royalties from traditionally translated royalties in some draft language. Several agents report that MT-disclosure language is now among the top five negotiation points on translation deals.
Translators’ associations, for their part, are pushing post-editing rate cards and refusing work pitched at machine-output pay. The market is listening: a growing share of rights departments now ask bidders to declare their workflow, and a small but real premium is emerging for certified human-only translation.
The divide will not close soon. But 2026 may be remembered as the year the industry stopped debating whether the technology exists — and started negotiating who controls it.