Pixel International Conferences

Speech recognition has reached a level of accuracy where it is  powering automatic translation and testing. What impact has it on language teaching? How should we develop appropriate pedagogical models  and prepare teachers for its  application to our classrooms?   I will give a critical analysis of its pedagogical uses and dangers.

The technology of Automated Speech Recognition (ASR) is rapidly becoming more sophisticated and it is becoming part of everyday life.

Earlier applications to language teaching were very inaccurate and did not aid learning. But the current accuracy levels of ASR mean that this is changing.

Speech recognition facilitates automatic translation. There are already mobile apps that allow students to speak into a phone or tablet and instantly hear the spoken translation.

These 'speech-to-speech' systems are mainly accurate in narrow domains (eg domestic or tourist language) but are likely to impact on students' motivation and expectations of learning languages.

ASR facilitates computer-based automatic marking of language teaching examinations - both written and spoken exams.  Cambridge University has set up a new institute, ALTA, to research this and is trialling the automatic marking of Cambridge English language exams.

It also facilitates auto-response to communicative interactions in the classroom, where students can use their tablets (in pairs) to speak or write responses to a task and get an instant correction or formative assessment.

It also facilitates new ways to work on phonology and accent - using IBM's programme 'Reading Companion', for example.

I will address these questions:

What is the impact of this for teachers in the classroom?

What is the impact on teachers’ need for training and development to be able to use this technology in the classroom and adapt to its use in examinations?

I will give a critical  appraisal of the application of Speech Recognition  to language teaching and consider the impact on pedagogy and  teacher development needs this may entail.