MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641073678 A) filed by Asma Jabeen, Assistant Professor Department Of English, Methodist College Of Engineering & Technology.; Dr. C. S. Srinivas, Assistant Professor Of English Department Of M&h, Mahatma Gandhi Institute Of Technology A.; Dr. Bhakti Natoo, Assistant Professor Department Of English, Vardhaman College Of Engineering.; Dr. Kesharaju Nirmala, Professor Of English Department Of H&s, Keshav Memorial College Of Engineering.; and Dr. Nidhi Mishra, Associate Professor Department Of English H&s on June 13, 2026, for A Data Centric Approach To Evaluating Generative Ai Chatbots For English Speaking Practice.
Inventors include Asma Jabeen, Assistant Professor Department Of English, Methodist; Dr. C. S. Srinivas, Assistant Professor Of English Department Of M&h; Dr. Bhakti Natoo, Assistant Professor Department Of English; Dr. Kesharaju Nirmala, Professor Of English Department Of H&s, Keshav Memorial College Of Engineering.; and Dr. Nidhi Mishra, Associate Professor Department Of English H&s, Gokaraju Rangaraju Institute Of Engineering & Technology..
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: Abstract The present invention discloses a data-centric evaluation framework the GenAI-ESP Evaluation Framework for the systematic, reproducible, and multi dimensional assessment of Generative AI chatbots deployed in English speaking practice contexts. The framework integrates a five stage data pipeline comprising: standardised speaking Task Design calibrated to CEFR proficiency levels; Interaction Capture of learner chatbot exchanges; Automated Feature Extraction computing six quantitative metrics Pronunciation Accuracy Score (PAS), Grammar Error Detection Rate (GEDR), Lexical Diversity Index (LDI), Response Coherence Score (RCS), Conversational Fluency Index (CFI), and Adaptive Feedback Quality Score (AFQS); a Pedagogical Assessment Rubric (PAR) evaluating chatbot responses across six pedagogically validated dimensions; and a Learner Outcome Measurement protocol assessing pre post speaking proficiency and self-efficacy. A Comparative Benchmarking Module computes Composite Evaluation Scores for multiple AI chatbot platforms. The framework addresses the critical gap in AI assisted language education evaluation by providing a standardized, scalable, and pedagogically grounded methodology for assessing Generative AI chatbot suitability for English speaking instruction.
Disclaimer: Curated by HT Syndication.
