Editors launch 'Don't ask AI, ask a peer' seriesCEFR B1
15 Apr 2026
Adapted from Ameya Nagarajan, Global Voices • CC BY 3.0
Photo by Marvin Meyer, Unsplash
Editors at three organisations co-wrote an editorial to confront the growing use of large learning models known as AI. The editorial was written by Ameya Nagarajan, Débora Prado and Hija Kamran and was only copy-edited for punctuation and spelling to preserve their distinct voices. APC and GenderIT.org developed an AI policy last year with inspiration from Global Voices; they discussed human voice, agency, creativity and labour while considering how to respond.
From those discussions came a collaborative series, Don’t ask AI, ask a peer, which aims to re-emphasise people-to-people knowledge sharing. The editors say generative AI has reduced such exchanges as people turn to automated tools instead of asking family, elders, coworkers and community members.
The series began with two questions posed like prompts to AI: why value human creativity and how to create a human-rights approach to AI. Authors and illustrators from different countries will answer. Content will be edited and translated by people, include human emotions and contradictions, be remunerated, and list clear acknowledgments. The editors invite peer exchange, feedback and debate; the series is published throughout April on APC.org, GenderIT.org and globalvoices.org.
Difficult words
- editorial — short article expressing an organisation's opinion
- confront — face or deal with something difficult
- generative — able to produce new text or images
- agency — ability to make independent choices or decisions
- remunerate — pay someone for work or servicesremunerated
- peer — person of similar age or status
- acknowledgment — statement that thanks or recognises helpacknowledgments
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- Do you prefer asking a person or using AI for advice? Why?
- How might paying authors and translators change the quality of articles in a series like this?
Related articles
New training method helps models do long multiplication
Researchers studied why modern language models fail at long multiplication and compared standard fine-tuning with an Implicit Chain of Thought (ICoT) method. ICoT models learned to store intermediate results and reached perfect accuracy.
Vietnam Rise supports grassroots activists inside Vietnam
Vietnam Rise, founded by members of the Vietnamese diaspora, supports grassroots activists in Vietnam with fellowships, grants and regional cooperation. The group faces state pressure and works to raise international awareness.