Embedding translation AI in the L2 secondary school classroom: creative applications and potential barriers

Nick Riches

Automated translation is becoming ever powerful, with sophisticated platforms such as Google Translate increasingly taking work from human translators. Translation has also become important within the school language syllabus post 1916 due to government driven changes in curricula. However, automated translation has yet to play an important role in the language classroom.  This study investigated teachers' attitudes to automated translation, and simultaneously created an automated translation app for classroom use, which combined the power of Google Translate with a user-friendly interface, designed using open source tools. Furthermore, the app coloured-in word classes, e.g. Nouns and Verbs, with a view to demonstrating sentence structure.

black text on a white background

An early version of the app was developed and used as discussion point in a focus group with language teachers working in local schools. Data from the focus group were analysed using Brown and Clark's Thematic Analysis following the steps of familiarizing with the data through transcription and re-reading, initial coding, collating codes, formulating and refining the themes and the generation of clear definitions of the final themes.

Teachers' attitudes towards translation technology were mixed. On the negative side they were concerned about using translation technology for cheating, especially when students were required to conduct an actual translation exercise. There were also issues around the use of technology in the classroom, e.g. the need to keep students focused on a particular task, particularly with open-ended devices, e.g. those containing multiple apps, or with browser capability. On the plus side, participants were impressed with recent improvements in the accuracy of online translation and felt that the pilot app provided the kind of "instant gratification" which would keep students motivated. They thought that the app provides "a way of harnessing google to actually teach them how the languages work ". They were keen on the ability to analyse the structure of sentences using colours, and compare the sentences on the page to visualise differences in structure. One of the participants felt that this would be particularly useful for students learning English as an additional language if translation were possible from their first language. One teacher had already employed a methodology comparing the structure of languages which they labelled "dodgy English". This involved translating sentences word for word in order to highlight differences across languages, a technique which is similar to the way the app works. Another mentioned the usefulness of an app which translates between multiple languages:

“because err I have Spanish people that want to learn Greek I have English that- that know other languages like German and other stuff so they’re trying to do the comparison themselves so I don’t think- I mean I would like to give them but I just know English and Greek and some French so I’m not very good at it. But if an app could do that hehe like for me.”

Based on the focus group, the app was developed further, with the most recent version available here, along with the computer code. The app incorporates the teachers' suggestions to provide control over colours, and the ability to focus in on particular sets of word classes, e.g. nouns or verbs. The code for the app is freely available so any school may use the app and host it, though there are a couple of limiting factors, namely the need to host it on a server, and the need to purchase a key to use Google Translate. Nonetheless this latter factor should not be a limitation as translation costs are generally low, and educational licences may be obtained.

Further work on the study, which has been partially disrupted by COVID, will involve visiting schools, presenting the app to school teachers, and eliciting further feedback within a group setting.

The "Embedding translation AI in the L2 secondary school classroom: creative applications and potential barriers" project research committee includes Nick Riches, Müge Satar Cohen, Elaine Lopez, and Saziye Tasmedir.


Dr Nick Riches (@NickRiches1 on Twitter) is Senior Lecturer in Speech & Language Pathology at the School of Education, Communication, and Language Science at Newcastle University. His research is interested in “linguistic” approaches to Specific Language Impairment (SLI).