Carnegie Learning’s MATHstream: Merging the hottest trends in tech to engage math students

Born out of Carnegie Mellon University 25 years ago, Carnegie Learning has been on the cusp of cutting-edge educational technology for nearly a quarter-century, fine-tuning products to help students learn mathematics best via years of data analytics and software improvements.
It’s no surprise that the group’s latest tool, made possible via support from the Learning Engineering Virtual Institute (LEVI), combines three of today’s hottest trends in tech – influencers, streaming, and generative AI – to support mathematics teachers while engaging middle school students. The result, MATHstream, is a personalized math video platform for grades 6-12. The adaptive program includes a library of streamable videos which are accompanied by AI-powered avatars that can nudge students with questions, hints, and rewards. The tool also includes “stream statistics” of skill assessments that track students’ performance on assignments.
Overall-MATHstream
Math teacher and influencer Robert Ahdoot and a student watching a video of Robert explaining a math concept in MATHstream.

In building MATHstream’s library of streamable math videos, Carnegie Learning scoured platforms like Instagram, TikTok, and YouTube for today’s best on-camera teachers and influencer educators. The group then partnered with them to create high-quality student-friendly videos. MATHstream features Mrs. Deidre Kelly, a 6th grade math teacher from Texas known to her 1.4 million TikTok followers as “@the_mrskelly,” Robert Ahdoot, a math teacher and the founder of online math platform YayMath.org, and Howie Hua, a college math instructor at Fresno state and TikTok and YouTube extraordinaire.

It has proven to be an effective and engaging way to offer AI-based tutoring, by pairing engaging videos from Kelly, Ahdoot, and Hua with real-time questions, prompts, and feedback. This combination can be especially helpful for low-income classroom settings that often lack access to traditional means of tutoring and classroom support. The game-based learning program delivers questions and videos from the streamers to the students. If students answer correctly, they proceed to the next set of questions and videos. But when students make an error, the AI tutor comes into play with adaptive suggestions, feedback, and instruction.
“When we’re using MATHstream we’re able to see a student struggling on a certain lesson, they get immediate support, and then as a teacher, you’re able to help them with that lesson they were struggling with,” MATHstreamer Mrs. Deidre Kelly said in a video while explaining the tool.
Via detectors, the tool can also gauge if a student is bored, confused, frustrated, or gaming the system – in response to which the tutor can also suggest mindfulness exercises such as deep breathing.

Overall, MATHstream can help scale tutoring accessibility, combat teacher fatigue and strain as a result of the ongoing teacher shortage, and aim to dramatically improve outcomes for struggling students.Carnegie Learning’s initial pilots of MATHstream had over 2,200 active learners by the end of the 2022-2023 school year. What’s more, a larger catalog of content is now available, and an additional 2,279 students have been assigned content, while 7,815 students have committed to MATHstream trials. By 2027, Carnegie Learning aims to improve math outcomes for approximately 447,500 students using MATHstream.

building-MATHstream
“Math software is flat because of significant amounts of text students have to read,” said Carnegie Learning’s Jamie Sterling. “This is an amazing example of using generative AI for good in the classroom because it showcases the best of learning science research, and how we can personalize for students in a way that has never been possible before. We can create content based on students’ interest on the fly, and bring it to life for them.”