Magpie Literacy*

*ALTERNATIVE PROFILE

‍

We included Magpie Literacy in our tour based on public information about the product roadmap. On the tour, we learned that, though Magpie is testing AI-powered features for future development, the current product does not include any AI-features (i.e., they only use deterministic algorithms developed by humans; they do not use generative AI, predictive AI, or automatic speech recognition technology in the current, standard version of the product).  

‍

We observed Magpie in schools and interviewed users using multiple versions of Magpie’s offerings, including some with beta features of AI elements—specifically automatic speech recognition, which Magpie may include in its standard product offering once the features are accurate enough to provide instructional value to students.

‍

Because 1) this tour is focused on AI-powered features and their benefits and risks, and 2) Magpie does not currently include AI in its standard product offering, we are not doing a full product profile in this round of the tour. However, below, we documented our understanding of how Magpie is considering using AI as well as our instructional take on the wins and limitations of those uses. If AI-powered elements become part of the core Magpie offering, we will re-review and publish a full profile of the product. 

‍

Here is what we know about Magpie and the ways AI may play a role in its future offerings:

‍

What Magpie’s current offering looks like in action

• What the students do: Students work independently on their device. They watch a direct instruction video led by a human, on-screen teacher, then move through skill games: naming letters and sounds, identifying beginning, middle, and ending sounds, blending and spelling with mouth-movement and voice models, and reading words and then short sentences. This includes isolated practice, integrated skill practice and applying in text. Students get multiple attempts at a task and practice each skill across different task types before moving into reading words and sentences that use it.
  ‍
• What the tool does: Magpie runs the assessment and guides the student's movement through the skill sequence. It marks each response right or wrong, and, when a response is wrong, it tells the student the answer is not right yet and opens another try. After a second wrong try, it shows the correct letter or mouth positions and lets the student attempt it again. The assessment screens and monitors foundational skills over time, feeding results into what the student works on next and into the teacher's reports.   
  ‍
• What the teacher does: Teachers facilitate students' independent skills practice. From there, they can monitor the dashboard, which reports lessons completed, skills practiced, time on tool, and assessment benchmark results, to see whether students are pacing appropriately and staying engaged. They can also connect the isolated skills students practice on Magpie to the reading, writing, and comprehension work in their core curriculum.
  ‍
• What the leaders do: Leaders adopt Magpie and set the scope of where it is used, such as whole-class Tier 1 foundational instruction or intervention and enrichment blocks. They can review usage, completion, and skills data to support teachers in targeting small-group instruction.

‍

Product development and AI-powered features in testing/under consideration

• Magpie Literacy is currently testing and evaluating automatic speech recognition technology (which does not involve generative AI but does include machine learning which is a kind of AI) to support stronger assessment of student reading and provide students more feedback on practice.

‍

Instructional wins and limitations with automatic speech recognition technology

• Wins: With the speech recognition features enabled, students can get more precise feedback on their letter-sound combinations during station rotations, and teachers can get more targeted information about students reading strengths and areas of need from authentic practice rather than just multiple choice questions and exercises. 
  ‍
• Limitations: Speech recognition software has been found to be as good at understanding students as humans; however, teachers have cited that dialect variation can cause the software to misunderstand student responses, so teachers may need to continue to monitor responses to ensure accuracy.

‍

Learn more

• Instruction Partners' CEO Emily Freitag spoke with Rebecca Kockler from Magpie to understand what problems they're trying to solve, how the tool works, and what they're still figuring out. Watch their conversation here.
  ‍
• Visit the Magpie website to learn more: https://www.magpie.org/