C-LARA

Overview Use C-LARA Example content Ongoing Activities Contributors Documents Technical issues Blog and Discord Contact 2024 workshop

C-LARA: ChatGPT-based Learning And Reading Assistant

C-LARA, or ChatGPT-based Learning And Reading Assistant, is an open source web platform, currently under development by an international open source consortium, whose purpose is to make it easy to create engaging multimedia texts that can help language learners develop their reading and listening skills. The platform draws inspiration from the earlier LARA project but has been completely rewritten.

As the name suggests, ChatGPT-4 is at the centre of our activities. It is used in two complementary ways, both as a software component and as a software engineer. In its software component role, ChatGPT-4 performs all the core language processing functions in the platform. It can write texts, divide them into sentences and lexical units, identify multi-word expressions, add glosses in another language, and annotate words with lemmas (root form) information. Audio is automatically added for the wide range of languages where a TTS engine is available.

As a software engineer, ChatGPT-4 has played a central role in creating the codebase, which is freely accessible on GitHub. ChatGPT-4 has produced about 90% of the code, working in close collaboration with its human partner Manny Rayner. It is also responsible for the greater part of the software design, and has written a substantial portion of the academic publications which the project has produced. We consider that ChatGPT-4 has amply demonstrated its right to be accepted as a core member of the project; our policy of consistently crediting it as a coauthor has already led to some controversy.