CorpusMate combines artificial intelligence with 47 million words of authentic language data, making evidence-based language learning and research more accessible for teachers, students, researchers and language learners.
What's the problem?
Corpus linguistics uses millions of examples of real-world language to reveal how people write, speak and communicate in different settings. These insights can improve language learning, teaching and research, but the tools used to access them are often complicated and designed for specialists.
This means many teachers, students and language learners cannot easily benefit from evidence-based language insights and instead rely on general language advice or guesswork.
What's UQ's innovation?
Developed by Associate Professor Peter Crosthwaite from UQ's School of Languages and Cultures, CorpusMate combines AI with large-scale language datasets to make corpus linguistics easier to use and understand.
The platform draws on 47 million words of real-world text and speech from academic writing, scientific research, lectures, news, digital media and everyday conversation across more than 20 subject areas.
At the heart of CorpusMate is an AI assistant that helps users make sense of what they find. Rather than simply generating answers, it analyses search results and explains patterns in vocabulary, grammar and usage based on authentic evidence. It can help users explore how communication changes across different contexts, create learning activities and support research and writing.
CorpusMate adapts its support for researchers, teachers, students and language learners, and provides responses in 11 languages, making evidence-based language analysis accessible to a much wider audience.
What's the impact?
CorpusMate is making evidence-based language learning and research accessible to a much wider audience.
Since its public release, the platform has been adopted in language teacher training programs across Asia-Pacific and Europe and has attracted more than 1,000 users since April 2026. Feedback from more than 180 users has helped shape the development of CorpusMate 2.0, ensuring it meets the needs of researchers, teachers, students and language learners.
By combining AI with millions of words of authentic language data, CorpusMate helps people discover how language is really used. Instead of relying on generic AI-generated advice, users can explore real examples, identify patterns and gain evidence-based insights to support learning, teaching and research.
Did you know?
Unlike many AI tools, CorpusMate does not generate language advice from scratch. Instead, its AI assistant analyses a user's actual corpus search results and provides explanations grounded in authentic language evidence.
Led by
- Associate Professor Peter Crosthwaite
Deputy Head of School, Applied Linguistics Program
School of Languages and Cultures
| AI Research Strength | Human-Centred AI |
|---|---|
| Industry Portfolio | Social & Economic policy |
| Key Publication | Crosthwaite, P. (2026). CorpusMate 2.0: Integrating AI-powered language analysis into a data-driven learning corpus tool. Artificial Intelligence in Language Education, 2, 104826. https://doi.org/10.29140/aile.2026.104826 |