Skip to main content

Research team

ChitterChatter's research-led approach to AI speaking practice.

ChitterChatter's research work is led by Jadon Geathers, Co-founder and CEO of ChitterChatter and an Information Science PhD student at Cornell University.

The work focuses on AI-powered conversational agents for learning: systems that give students more authentic interaction, more repetitions, and more useful feedback without removing teacher judgment from the classroom.

Jadon Geathers
Jadon Geathers, Co-founder and CEO

Research lead

Jadon Geathers

Role
Co-founder and CEO, ChitterChatter
Affiliation
Information Science PhD student, Cornell University
Education
MS Computer Science and BS Mathematics, Stanford University
Research focus
Conversational AI, feedback systems, language learning, and scalable learning technologies
Location
Ithaca, NY

Jadon develops AI-powered conversational agents that support learning through authentic, personalized interaction. For ChitterChatter, that means designing speaking practice that is realistic enough to matter, structured enough for classrooms, and low-pressure enough for students to try, repair, and repeat.

At Cornell, Jadon is part of the Future of Learning Lab and is advised by Professor Rene F. Kizilcec. His ChitterChatter work connects learning-sciences research with practical instructor workflows: curriculum-aligned tasks, student work teachers can review, and feedback that helps students prepare for live communication.

Cornell profile · Jadon's research website

Research stance

Built around practice, confidence, and teacher interpretation.

ChitterChatter treats AI conversation as a way to increase meaningful speaking opportunities, not as a replacement for live instruction. The product is designed around language-learning ideas such as i+1, affective-filter reduction, Willingness to Communicate, role-play practice, and ACTFL-oriented communicative goals.

The practical design goal is direct: students should get more chances to speak in realistic situations, and instructors should get enough context to decide what needs modeling, follow-up, or assessment.

What the evidence shows

Research-informed practice. ChitterChatter gives students repeated speaking practice through teacher-created scenarios, with transcripts, recordings, and feedback to support review. Our methodology explains the learning research behind these choices.

Published ChitterChatter research. Jadon Geathers, A. J. Alvero, and René F. Kizilcec authored ChitterChatter: Curriculum-Aligned AI Speaking Partners for Language Learning Classrooms for the 2025 ACM Learning @ Scale conference. Explore more work from the Future of Learning Lab.

Early classroom experience. Students in three college Spanish classes piloted ChitterChatter. In interviews, students described feeling more comfortable practicing with an AI partner and preferred ungraded practice. Read about their experience in Patricia Waldron's July 2025 Cornell Chronicle article.

What we are still learning. The pilot offers insight into students' experience, but does not establish improved proficiency, long-term retention, or better results than other practice methods. Results may differ by learner and language. Instructors should review AI feedback and student work to decide how ChitterChatter fits their course.

News and recognition

Related reading