AWS Public Sector Blog
EdTech innovator Sibme, powered by AWS, provides educators with AI-based instructional feedback
As a teacher with the KIPP charter school network in Houston, Texas, Dave Wakefield knew there had to be a better way for educators to gain insightful feedback on their instruction.
Traditionally, educators who wanted feedback on their teaching either had to have someone visit their classroom or film themselves and then send that video to a mentor or peer for review. The process was cumbersome, requiring teachers to upload videos onto a shareable drive and share comments haphazardly with one another through chat apps, collaborative documents, or email. Insights, while well-intentioned, were not actionable, and when feedback was finally delivered to the instructor, tracking progress over time was a challenge. It wasn’t always apparent which changes to instructional methods should be credited for the gains.
In 2013, Wakefield founded education technology (EdTech) company Sibme as a way to use technology, powered by Amazon Web Services (AWS), to help educators access quicker and more reliable feedback.
Bringing the power of AI to educators
As Wakefield and the Sibme team were exploring ways to improve the teacher-feedback process, new artificial intelligence (AI) tools were entering the marketplace. Jeff Ritter, Sibme’s CIO and a former teacher and school administrator, realized the implications of this groundbreaking technology while teaching computer science classes.
“For several years, I had been teaching a segment on AI and had students delve into questions like, ‘how do you train an AI model?’ And, ‘what can AI do for us?’ We immediately realized that AI offered a great way to obtain coaching or feedback,” said Ritter.
A former instructor himself, Ritter saw that AI could be used to offer teachers private feedback on the qualities they were typically evaluated on—which had been largely subjective in the past. For instructors, things like teacher talking time, questions from students, student engagement, sentiment, and even their movements within the classroom were both critical to teaching and were difficult and time-consuming to measure. Empowering AI to do more of the measuring could make these qualities quantitative, providing concrete feedback for teachers.
AWS provides the building blocks for Sibme’s AI solution
The first step to building Sibme’s new solution was finding a way to use AI to transcribe instructor videos. To do so, Ritter’s team worked hand-in-hand with engineers from AWS to ensure everything worked perfectly.
“Since what the AI is tracking—classroom behavior—is all qualitative, it’s incumbent on the tools that we use to give us great data back,” said Ritter. “Our entire solution is based on these transcriptions being very accurate, and that’s where AWS really stepped up.”
Far more than an out-of-the-box transcription solution, AWS proved they were willing to get hands-on and provide personalized help. When an issue with speaker diarization (distinguishing between different speakers) was preventing the identification of unique voices within a classroom, for example, the AWS account team worked swiftly with the product team to get the problem resolved.
“This was a great example of how responsive AWS was,” said Ritter. “I was able to immediately tell our development team, ‘OK, it’s fixed.’” The result worked so well that it was pushed out as a product-wide software update for all Amazon Transcribe users.
Based on the high level of both success and service, Sibme decided to continue building on AWS, eventually adding both Amazon Rekognition and Amazon Comprehend to their technology solution. And since security and compliance are especially important for schools, where there are rules around data security and the recording of students, Sibme appreciated the AWS toolkit that helps them stay FERPA and COPPA compliant while securing data and shutting down suspicious activity before it can compromise accounts.
“AWS connects you with the right people, helps you brainstorm, and introduces new resources and solutions as you grow,” said Ritter.
Helping educators track progress
Unlike other apps available in the education industry, “The unique thing about Sibme is that it allows the user to attach evidence to show growth on what they’re working on,” said Ritter.
In order to increase engagement and hands-on learning among students, for instance, many educators work to reduce their teacher talking time. With Sibme, an instructor could show evidence that at the beginning of the year, they spent 80 percent of the time talking. By midyear, they had reduced that to 72 percent, and by the end of the year, they were at 58 percent––close to their goal of 50 percent.
“Teachers love to show that they’ve made progress, and that’s where AI and that quantitative data really can help,” said Ritter. “It’s powerful to have evidence tracking throughout the year to demonstrate that progress.”
Teachers can also curate what they share with the tool, and they can run the program on their own videos—at any time—for self-reflection and improvement. When they are ready to share, the platform enables synchronous feedback through timestamps so that everyone is on the same page.
Even teachers who already have formal coaching resources benefit from Sibme’s tools since their mentors are now equipped with data, definable metrics, and evidence of growth. The end result? Discussions around teacher progress become fruitful, targeted, and tangible because less time is wasted discussing whether something did or did not happen in the classroom.
Moving beyond the classroom
Sibme’s tools focus on helping individual educators, but when implemented at scale, the entire school is improved. “Texas is our home state,” said Ritter. “Here, schools get ranked, and if you’re below a certain level, you’re failing. We found schools that have started using Sibme have moved from a D to a B-minus rating.”
The traditional school setting is only a starting point for the team, however. Sibme has plans for an AI-assisted chat to help teachers learn and adjust in real time. This technology can be used in multiple locations, such as the football field or even board meetings.
Sibme’s reporting and analytics will expand to include more functionality, as well. Look for new features (including a formative assessment), reporting that documents how well a teacher followed their lesson plan, the ability to have a conversation with AI using the documents and comments made on the video, and recommendations for how to actively engage this new generation of students. The platform also hopes to make searchability a core feature, giving educators a way to easily find the number of videos uploaded in a month or their average wait time over the past two weeks.
As AI technology expands, so will the promise that Sibme offers its educator community. The AWS partnership is central to delivering on the promise, and both companies look forward to seeing how the rise of AI builds upon traditional pedagogy to empower teachers and create lasting results in the classroom.
Learn more about AWS and its commitment to education technology on the AWS for Education page.