How AI can deepen inequities for non-native English speakers in science

A paper co-authored by Tatsuya Amano was rejected recently without review because its level of English did not meet the journal’s required standard. His research suggests that 38% of researchers who are not fluent in English have experienced similar rejections.

Amano, whose first language is Japanese, describes how dismantling language barriers will result in improved knowledge sharing, and in the long run, better research.

Journals, he argues, can help by taking steps to distinguish the quality of science from the quality of language when assessing manuscripts. And conference organizers can adopt a range of measures to support presenters and attendees whose first language is not English.

The biodiversity researcher is one of eleven scientists leading TranslatE, a project which strives to make environmental science more accessible to non-fluent English speakers.

AI and translation tools can bring huge benefits to researchers like him, he says, but they won’t all have been trained on many of the world’s estimated 7000 different languages, deepening inequities in science. Cost is another factor, particularly for those in global south countries. “People from high income countries may be more likely to benefit from those emerging AI technologies,” he says.

Hosted on Acast. See acast.com/privacy for more information.