Submit Here

Meet the 15 Finalists for the Dataset Prize | 2025 Tools Competition

We are excited to announce the 15 finalists named for the Dataset Prize of the 2025 Tools Competition!

In this track, 77 teams sought to address the field’s need for more high-quality public assets to advance learning engineering as a whole. Winners will receive $50,000 to prepare and release robust education-focused datasets.

Teams in this track are addressing the critical need to make innovation and research related to education less expensive and more inclusive. Their work will also support the field in responding to questions of privacy and bias that remain in the use of AI in education.

Finalists of the Dataset Prize

ALPS will collect fine-grained student interactions on a variety of computer science activities to understand how students learn with an AI-native education platform that offers a high degree of personalization, based on a learner model further enhanced with LLM-generated content.

AI-FIT will leverage AI-driven speech recognition and natural language processing to collect student reading fluency data—enabling educators to assess, personalize, and improve fluency for students with dyslexia and reading challenges in grades 3-5.

AI Model-based Tutor in STEM will collect facial action unit data to help understand students’ emotional and cognitive states from facial expressions during AI math tutoring.

BLAST will collect bilingual speech data from students in grades 3-5 to better inform performance standards for assessing and supporting bi/multilingual learners.

Bookbot will collect oral responses and learner data to understand literacy, thinking, and socio-emotional development across diverse languages and settings.

Phonologie will collect diverse, multilingual speech datasets from 5,000 children to improve the accuracy of Automatic Sound Recognition systems—allowing students to be better served by edtech products that incorporate speech recognition.

  • English Test for All (ETFA) | English Quest (India)

EFTA will collect English assessment scores, including audio recordings, of 10,000 students in three Indian states on a baseline and endline test 12 months later—generating an AI training dataset that will also help LLMs understand Indian children’s accents.

Excuela collects learning engagement, behavioral insights, and workforce upskilling data from frontline workers and vulnerable populations to understand how MSMEs can increase productivity levels, efficiently adopt circular economy practices, and drive sustainable impact.

  • High Quality Feedback for Essay Revision | Stanford University (United States)

High Quality Feedback for Essay Revision will collect written in-line feedback from experienced ELA teachers on students’ persuasive and literary analysis essays—identifying research-based dimensions of high quality feedback to observe the pedagogical characteristics of feedback.

Journify Learning will collect a wide range of data (e.g. student characteristics, student performance over time, IEP data) from various sources (i.e., across support providers), to help drive outcomes for students with special needs.

Mathlets will collect data from tens of thousands of middle school students to inform whether heterogeneity in the way students exhibit math anxiety, including those with learning differences, can be identified and alleviated through edtech tools’ use of personalized graphics, simulated instructional responses, and human-crafted dialogues.

Factors Education’s dataset will collect semi-structured chat data between secondary education students and an LLM assistant to benchmark the AI’s ability to identify and extract relevant learner profiles from open conversations.

SocialWise will collect data from autistic adults over the course of a college semester via survey and roleplays/chats with an AI agent designed to improve their communication and relationship skills—providing insight into the meaningful challenges of autistic college students, their ability for skill improvement, and impact on social and academic success.

SkillWise will gather performance and behavioral data from underserved communities to better understand how self-motivated learning systems can be a positive force for changing life trajectories. 

Learning Equality will collect anonymized search query data from its offline learning platform, Kolibri, to understand user content needs and enhance search and recommendation algorithms for underserved learners.


See finalists for all tracks here.

The Tools Competition has three phases and multiple prize levels to help winners create or enhance their tools at all phases of development. As they enter the third and final phase, these finalists will provide final evidence to support the development of their education-focused dataset. 

The Tools Competition has previously named 130 winners from 44 countries—projected to impact 131+ million learners worldwide, from early childhood to adulthood. Winners for all opportunities in the 2025 competition cycle will be announced in summer 2025.

The 2025 Tools Competition is a program of Renaissance Philanthropy and is supported by: Schmidt Futures, Griffin Catalyst, Walton Family Foundation, Axim Collaborative, Oak Foundation, and Calbright College.

Want to be kept in the loop for the next competition cycle? Join our mailing list!

RECENT POSTS

Back to all posts
Announcing the Finalists of the 2025 Tools Competition!