Back to Research
Too often, America’s public schools are forced to wait for a child to fail before offering the reading intervention they need to succeed. This isn’t due to a lack of teacher empathy or expertise, but rather a systemic resource gap that robs educators of the time and real-time insights required to step in early.
To break this cycle, teachers need a new generation of tools—a technological radar that can instantly pinpoint exactly where a K-2 reader is struggling and immediately provide personalized strategies to get them back on track.
This may seem far from today’s reality, especially at a time when nearly 7 out of 10 fourth grade students are not proficient in reading. However, Renaissance Philanthropy is placing a major bet on rapid advancements in artificial intelligence to bridge the gap between the earliest signs of a reading struggle and targeted classroom instruction. Accelerated AI—from automatic speech recognition to continuous data loops—can do far more than just supplement teaching; it can add a hyper-precise diagnostic layer that identifies why a child is tripping over a word and immediately charts their path toward progress.
To make this universal support a reality, Renaissance’s Learning Engineering Virtual Institute (LEVI) Literacy has launched a five-year moonshot initiative with a clear, measurable goal: to build low-cost, scalable tools that can halve the number of K-2 readers falling behind grade level.
LEVI was designed using best practices from successful R&D programs like DARPA. Rather than funding isolated pilots, the LEVI model centers on a clear, ambitious “North Star” goal. We match that ambition with significant multi-year, financial investments and embed our grantees within a cross-cutting learning community. This method proved its power in our inaugural LEVI Math program, and I am convinced it is the fastest way to drive equitable, low-cost, and scalable solutions across public education.
Meet The LEVI Literacy Cohort
Four teams have been selected to take on the moonshot goal:
ProjectRead AI: Building on the highly successful University of Florida Literacy Institute phonics curriculum, ProjectRead is developing a student-facing AI phonics tutor using phoneme-level Automatic Speech Recognition (ASR). This tool connects a teacher planning portal directly to the student’s digital workspace, creating a seamless feedback loop where a child’s specific reading gaps immediately shape their next personalized practice session.
LitLab: LitLab offers a curriculum-agnostic platform where students read decodable texts aligned to their district’s phonics sequence, generating continuous mastery data as a byproduct of natural practice. LitLab will leverage LEVI funding to build a neural network-based mastery model using advanced knowledge tracing, while releasing an open-mastery API as a shared public good for the edtech field.
OxEd: Defying the standard edtech playbook, OxEd focuses on oral language, an essential but under-targeted component of reading comprehension. OxEd uses a screen-free approach where AI and ASR are entirely teacher-facing. The platform provides educators with real-time insights and longitudinal data to deliver high-fidelity, targeted oral language interventions directly to K-2 students.
Harvard READS Lab (MORE TEAMS): The MORE TEAMS model (Teachers Engaging and Motivating Students) targets the often-overlooked areas of vocabulary, background content knowledge, and text comprehension. The project uses AI to listen to classroom-level teacher interactions to improve instruction, coupled with a student-facing app that generates personalized content in subjects like science and social studies.
