Artificial Intelligence: What Schools Can Actually Learn From Their Data

Establishing Data Governance Policies

As a heavily regulated industry, K–12 education requires districts to comply with federal and state policies to protect personal information and students’ online activity. Microsoft has an education AI toolkit to help schools navigate their data governance policy creation by “emphasizing transparency, privacy, security, human oversight and accountability,” explains Jubelirer. “A strong data governance policy begins with purpose.”

Starting with identifying the exact problem to be solved, Jubelirer recommends districts follow standard data governance policy best practices:

• Collect only the information needed for educational purposes
• Limit access to authorized staff only
• Use aggregated or de-identified data wherever possible
• Maintain strong security controls
• Be transparent with families about how data is used

Eric Crespo, superintendent of Weehawken Township School District in New Jersey, says his district’s data governance policy is “the rulebook for who gets to see student data, how it’s stored and which vendors are allowed to interact with it.” He adds that approved vendors are given a “key to exactly one room in the house, not the whole house.”

In Weehawken Township, the data governance policy was formed with the assistance of a law firm and then approved by multiple levels of stakeholders. “It was a mandatory state regulation setting the floor, a technology administrator filling in the district-specific details and a public board vote making it official,” explains Crespo.

This policy also governs the entire lifecycle of student records in Weehawken. Students and their parents must fill out an authorization form before accessing the district’s network for the first time.

Addressing Data Bias

Every school has its own culture and context, all of which needs to be taken into consideration when reviewing the insights that AI-powered data reveals.

 “AI-powered tools can help schools identify patterns, surface insights and analyze large data sets more effectively, but human expertise remains essential for interpreting results, asking the right questions and making informed decisions,” says Jubelirer. “The best outcomes happen when technology helps educators evaluate information more effectively while keeping people at the center of every decision.”

Crespo echoes the importance of human judgement, saying, “The first line of defense is refusing to let any single data point make a decision on its own.”

A student may have lost a loved one, or a fire drill may have taken place right before a failed test, Crespo explains. The data acts as a flag, pointing educators and administration to potential problems or inconsistencies; however, the context of the data and the face-to-face interactions with the child are how the decisions are truly made.

“Data is never allowed to make decisions on its own; instead, it acts more like a check engine light,” says Crespo. “We triangulate it, we disaggregate it, and, at the end of the day, a human being still has to look a kid in the eye and tell a student, ‘We believe in you,’ regardless of what the numbers say.”

Putting the Data to Work

Crespo’s district has achieved marked success in a few key areas since focusing on their data. He notes that the district ranked “higher than 81% of school districts nationwide in reading performance,” as well as in the 99th percentile nationwide for our rate of improvement in reading.” Crespo attributes these high marks to “adjusting instruction in the moment rather than waiting until the end of the year to see who fell behind.”

“People hear ‘data-driven’ and picture a spreadsheet making decisions in a back room, but that’s not really what it looks like day to day,” says Crespo. To him, it looks more like a teacher filling out a form because a student had a rough day; when a pattern emerges that indicates the student may be struggling, the school sends a specialist to the classroom. “The data doesn’t wait for June to tell us something’s wrong. It’s supposed to catch it while there’s still time to fix it.”

One area of improvement that Crespo is particularly proud of is being able to keep high-needs special education students in their own schools. Using the data, the district was able to prove it has safe, specialized programming for these students, saving the district hundreds of thousands of dollars per year in external tuition and transportation while also allowing these students to remain in their own schools. “That’s not just a budget line to us,” he says.

In the future, the district could expand into using data to help guide students on their career paths or identify mental health issues earlier.

“The districts getting real value out of data aren’t the ones with the fanciest dashboard. They’re the ones who’ve built the habit of actually looking, actually connecting the dots between departments, and actually acting on what they find instead of just archiving it,” says Crespo. “The districts that treat data like a compliance checkbox instead of a compass are going to be the ones left behind.”