What happened
Project LISTEN's Reading Tutor uses speech recognition to listen as children read aloud. It gives spoken or visual help when a student asks, pauses, gets stuck, skips a word or misreads one. In a seven-month randomized trial in two schools of an affluent suburban district, students in grades 1 to 4 used it in a supervised lab. They averaged 19 hours in sessions of about 20 minutes. The comparison group did sustained silent reading in class at the same time. Reading specialists chose the participants and left out the weakest readers. The What Works Clearinghouse (WWC) reviewed the trial, published in 2013, in June 2026. It found that the trial meets its standards without reservations, with 178 students analyzed. Users scored higher on a total reading composite (improvement index 27) and on a basic skills composite (38). The WWC rated both domains Tier 3, promising. Effects on comprehension, fluency and attitudes to reading were uncertain. The trial's authors report that students used the 2000-2001 version of the tutor.
Sources
-
Peer-reviewed study, by the makers.
Mostow, Jack, Jessica Nelson-Taylor and Joseph E. Beck. Computer-Guided Oral Reading versus Independent Practice: Comparison of Sustained Silent Reading to an Automated Reading Tutor That Listens. Journal of Educational Computing Research 49, no. 2 (2013): 249-276. ERIC EJ1076364, abstract as provided by the authors.
Eighty-eight students used the 2000-2001 version of Project LISTEN's Reading Tutor (www.cs.cmu.edu/~listen) in 10-computer labs, averaging 19 hours over the course of the year.
-
Assessment report, by authors with no tie to the system.
Literacy Achievement outcomes—Tier 3 (promising evidence) found for the domain
-
Descriptive report, by the makers.
Project LISTEN, Carnegie Mellon University. Project LISTEN Summary (the project's home page), last updated 5 June 2013.
Adapts Carnegie Mellon’s Sphinx speech recognizer to analyze the child’s oral reading
Details
AI system
- Reported configuration
-
2000-2001 version
- Workflow
-
Students read stories aloud to the tutor in a supervised computer lab for about 20 minutes a session. The tutor gave spoken or visual help when students asked, paused, got stuck, skipped a word or misread a word.
Follow-up
- Conditions for replication
-
A school-provided monitor supervises students in the lab. The trial excluded the weakest readers, who were with reading specialists during the sessions.
Entry details
- Sector
-
Education.
- Purpose
-
Project LISTEN built the Reading Tutor to give children individual practice in reading aloud, with a computer that listens and helps.
Not yet checked: model, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). A listening Reading Tutor beats silent reading in a trial (AIBD-0720). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0720.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)