ArXiv Will Require English Submissions - and Says AI Translators Are Fair Game

The preprint repository arXiv will require all submissions to be written in English or accompanied by a full English translation starting February 11, a policy change that explicitly permits the use of AI translators even as research suggests large language models remain inconsistent at the task.

Until now, authors only needed to submit an abstract in English. ArXiv hosts nearly 3 million preprints and receives more than 20,000 submissions monthly, though just 1% are in languages other than English.

Ralph Wijers, chair of arXiv's editorial advisory council, advises authors to verify any AI-generated translations. "Our own experience is that AI translation is good but not good enough," he says. A 2025 study from ByteDance Seed and Peking University ranked 20 LLMs on translation quality between Chinese and English; GPT-5-high scored nearly 77, just below the human expert benchmark of 80, but most models including GPT-4o, Claude 4, and Deepseek-V3 scored under 60.

Re:

[mistergrumpy]

ArXiv is support by donations. While anyone can submit, there is a moderation process. It's more about having an archival resource independent of the peer-reviewed journals.

Translation good enough

[cowdung]

If AI translation is good enough then readers should use that and leave the submissions in their original form.
International collaboration should be encouraged, not curtailed.

Re:

[ShanghaiBill]

If AI translation is good enough ...

Pure AI isn't good enough, and that isn't what most people will submit.

AI makes mistakes and humans make mistakes, but they make different mistakes. Humans make grammar and word choice mistakes much more often than AI, while AI makes factual mistakes.

A researcher can run the original through an LLM to produce a well-written, grammatically correct translation. Then they read it and correct the factual errors. Researchers who can't write English well can usually still read it well enough to catch factual erro

Re: Translation good enough

[Mr. Dollar Ton]

A researcher that can't write well will make the article unreadable again after correcting what they understand of the "AI" "translation", and they will likely miss the finer slop hallucinations.

The only way to correctly translate a scientific article is still by a human who actually understands both the subject and the two languages.

that's a mistake

[SG83]

the only person who claims that LLM are good for translation are people who don't know either topic or the language well enough. by submitting the translations only we make it impossible for translations to be improved later on. If we submit the originals, they can be later on translated either by humans or by the computers (or both).

Re:

[SG83]

Long time no see translation like that