One of the more intriguing new developments teased at the annual conference of the Consortium for School Networking, being held here this week:
Google is starting to provide cloud storage for K-12 districts, helping them warehouse the massive amounts of data they collect on their students, then offering artificial intelligence-as-a-service to help districts analyze the information and use it to generate visualizations, recommendations, and predictions.
The emerging effort involves Google’s and Data Studio tools. It’s currently being piloted by Washington’s Evergreen Public 69´«Ă˝ system and a couple other U.S. school districts, according to a presentation made here Tuesday.
During the session, Evergreen chief innovation officer Derrick Brown described his district’s partnership with the Mountain View, Calif.-based online-services giant.
“We have tons of data in our school districts,” Brown said, ticking off the information generated by student information systems, instructional software programs, online surveys of children’s social-emotional well-being, and special-needs students’ individualized education plans.
“All that data needs to go in a container. And that doesn’t exist. But this does that,” Brown said of Google Cloud Platform.
Google has not yet answered questions from Education Week about how the arrangement works.
But the basic idea, Evergreen officials said, is to take all the district’s information, put it in one place, then allow for automated analysis that can cut across datasets that have historically been housed in separate silos. Such capability has long been something of a holy grail for K-12 districts, which for years have pursued better data integration and interoperability by building their own information warehouses, providing learning-management and student-information systems, and pushing for new technical standards to ease the flow of data between systems.
It’s unclear exactly how Google proposes to overcome the longstanding challenges that have vexed other such efforts.
It’s also unclear how the new effort will address potentially significant privacy concerns, including questions over how participating districts will comply with the Family Educational Rights and Privacy Act. FERPA, as the nation’s main student-data-privacy law is commonly known, requires that districts maintain “direct control” of students’ educational records, and that parents be given the opportunity to review and amend those records.
“We haven’t gotten to the [data] ownership piece yet,” said Brown, who noted that Evergreen Public 69´«Ă˝â€™ involvement with the effort began under a since-departed superintendent and has proceeding without a formal data-sharing agreement in place.
“It’s something we have to address, as well as privacy,” he said. “Those are things we’re thinking about, but haven’t figured out yet.”
Evergreen recently starting hosting its data with Google, Brown said, as part of a free pilot experiment with the company. The district has not yet figured out what questions it hopes to answer with the data, and Google is not currently analyzing the data via its machine-learning algorithms or other artificially intelligent tools.
Big picture, cloud computing is nothing new in education. Eighty percent of districts now use cloud-based educational software, according to a
Amazon Web Services is already a popular option for K-12 cloud storage, minus the analytics and recommendations that Google is now exploring. And Google has been hard at work touting its cloud platform to higher education institutions.
But if the new Google experiment in K-12 were to take off--and overcome the many questions it is already raising--it could be a big deal.
Evergreen assistant superintendent Chris McMurray touted the effort as a way for schools to know more about their students, then combine that information with the powerful AI tools that Google is already using in its consumer products to personalize the learning experience.
“What I’m excited about is the ability to gather information on learners at such a scale that when I start seeing certain indicators, the machine can automatically [make predictions],” he said.
“Google knows me. It’s a great relationship, because Google can finish my sentences. We can harness that power to learn about our learners.”
See also: