Digging deeper into Lalilo’s Artificial Intelligence

Lalilo’s AI can determine which skills students have mastered and which they need to practice. It can provide them with work in specific areas while personalizing the level of content. It can help teachers by gathering enormous amounts of data and presenting it in a manageable and useful way so that teachers can ensure they are targeting the right skills and content for the right kids at the right time. Lalilo’s AI can help answer the questions of individualized instruction that plague teachers all over the world.

Let’s look at how Lalilo’s AI supports students.

In order to really understand how the AI of Lalilo works, you must first understand the structure of lessons, levels, and exercises. Lalilo’s phonics program is organized into lessons that represent an individual phonics component, for example, the short a sound or the consonant t. Each lesson is composed of various exercises, and each exercise includes content at a variety of levels. Within the consonant t lesson, for example, students can work on exercises that emphasize phonemic awareness, letter-sound correspondence, blending and segmenting, or letter recognition. Each of these exercises has content levels that help meet a student’s individual abilities within each skill. The graphic below helps to explain how the lessons, exercises, and levels interrelate.

When students log on to the Lalilo platform for the first time, they will have three lessons unlocked for them, which means the program can give them any exercise within these three lessons to begin their journey. Students begin with easier exercises, and if they are successful, they are given more difficult work. Student performance on the first exercises acts as the initial data that the AI uses to find the perfect match. Students work through a minimum of five exercises within one lesson. The level of the exercise is determined by how well the student performs on previous exercises. Once students demonstrate mastery of a lesson by achieving at least 80% success across all exercises, the lesson is “validated” and more lessons are unlocked. Up to seven lessons can be unlocked at one time if students are showing growth and getting close to validation. Students who show strength in lessons will only need to complete five exercises per lesson and will validate lessons quickly to unlock more lessons. Students who need more support will be assigned more exercises within a lesson until they are able to validate it. The more exercises a student completes, the more data the AI has to make informed decisions about which levels and exercises would fit a student best. All the student response data is put into the mathematical algorithm to assign each student the precise combination of lesson, exercise, and level that will fit his or her individual abilities and needs.

Lalilo’s AI is really complex! Let’s put this into a context to give you a better picture of how it might look for your students:

Sophia

Sophia is a strong kindergarten reader, knows her letter sounds and can already easily blend and segment CVC words. Sophia begins her Lalilo journey with a lesson on the letter t. She completes a phonemic awareness lesson, a letter-sound correspondence lesson, two blending lessons, and a letter recognition lesson. Sophia completes the exercises easily and receives a 90% success rate. Her letter t lesson is now validated, and she will not have to return to it. She unlocks another lesson, and now the program knows that she is strong in her phonics skills. In her next lesson, she will be given exercises at a higher level that will challenge her and keep her engaged in learning.

John

John is a emerging kindergarten reader. He is still developing his knowledge of letter-sound correspondence and is not able to consistently distinguish between similar looking letters. John is not yet able to blend CVC words. Just like Sophia, John begins his Lalilo journey with a lesson on the letter t. He completes the same exercises as Sophia, but he struggles with the lessons on blending and in letter-sound correspondence. John has a success rate of only 45% across the five exercises. The program will send John to another lesson (for example a lesson on letter m) and come back to the lesson on letter t later. The program will continue to give John exercises at an easier level until he is able to successfully complete them. Once John has reached a success rate of at least 80% on a lesson, it will be validated which will unlock other lessons.

Let’s look at how Lalilo’s AI supports teachers.

Lalilo’s AI can support teachers by simplifying and optimizing the processes of gathering and analyzing data. On the teacher dashboard, teachers are able to see which lessons their students have worked on and which lessons are unlocked. They can see the degree of mastery for each lesson for each student. On the dashboard, lessons in green have been validated, lessons in orange are ones students are still working on, and lessons in red indicate an area of high need. Teachers are able to see the success rates of different skills within each lesson. This information allows teachers to use Lalilo as a planning and preparation tool for whole class, small group, or individual lessons and assignments, saving teachers precious time in the planning process and ensuring that students are getting support in their areas of need.

What is the future of Lalilo’s AI?

Lalilo’s AI is still in its infancy. We are working everyday to improve our technology and to make Lalilo the tool that optimizes literacy learning and teaching. Here are some things we have in the works:

• The teacher dashboard will soon be able to group students by skills, levels, and lessons so that you can form groups easily and target skill gaps efficiently.
• We are working to suggest specific off-line lessons and activities that address the identified needs of a selected group of students for your small group instruction.
• We are working to create a instructional component (i.e. animated lesson) that is presented when the program detects a skill gap.
• We are teaching our AI to avoid too many of the same exercise back to back to increase student engagement.
• We are working on allowing teachers to input data (i.e. reading levels) and assigning skills, lessons, and levels correlated to the teacher-identified levels.

At Lalilo we value the incredible expertise and power that teachers have in facilitating educational experiences for children. Teachers have skills, knowledge, and an understanding of social elements that will never be replaced by technology. But we also know that having only one expert to address the individual needs of 24+ diverse learners is not realistic. Lalilo aims to become your personal teaching assistant, monitoring the progress of your students, giving you valuable data, and suggesting the lessons and activities that will target identified areas of need. Lalilo wants to be your teaching assistant to make your work as a teacher more efficient, more effective, and more personalized to the needs of your students.