Personalized Learning May Appear to Solve Pandemic Education Challenges but New Research Raises Big Concerns

Here we are at the beginning of July, and the COVID-19 numbers are soaring across many states. The recessionary collapse of state budgets makes public school funding look shaky, and nobody can quite say how to make school safe. With the 2020-21 school year expected to start in six weeks, school leaders face overwhelming uncertainty. It is easy to imagine that a well-known, free, online learning platform and curriculum, funded entirely by tech philanthropists, might be extremely tempting.

But there are reasons for caution. Last week, the Washington Post‘s Valerie Strauss published a guest column from researchers at the National Education Policy Center (NEPC), a guest column describing the Summit Learning Platform, and warning about some of the reasons for school district leaders to be cautious. These researchers, Alex Molnar and Faith Boninger, have documented serious concerns about protecting the privacy of student data collected by Summit Learning and its sponsors. To underscore the seriousness of their concerns, they describe the powerful people behind Summit Learning:

“Summit Public Schools, founded in 2003, currently operates 11 schools enrolling approximately 4,675 students in California and Washington State. In 2013-2014, there was no Summit Learning Program or ‘partner’ schools. Things changed quickly after a 2014-2015 agreement between Summit and Facebook established the goals of enhancing Summit’s self-created software platform and collaborating on a marketing strategy for its nationwide adoption. Since then, Summit Public Schools has received extensive technical support from Facebook and the Chan Zuckerberg Initiative and almost $200 million in funding from the Chan Zuckerberg Iniative, the Bill & Melinda Gates Foundation, and others.”

Molnar and Boninger review Summit’s marketing claims: “Summit Public Schools claims that its educational program does an exceptional job of preparing students for college and that its graduates succeed in college. Its specific claims are that its students are ‘100 percent Eligible for 4-Year College,’ that ’98 percent [of its students are] accepted to four-year college,’ and that its students graduate from college at ‘2X the national average.’… Schools that agree to become ‘partner’ schools agree to adopt Summit’s free, off-the-shelf program, use its digital platform, and participate in the required staff trainings and Summit-organized support activities… This marketing pitch has met with considerable success. By the 2018-2019 school year, nearly 400 schools nationwide, with nearly 3,800 educators and more than 72,000 students were using the Summit Learning Program.”

Molnar and Boninger’s report on their five years of research focuses on what ought to be serious data privacy concerns for school districts that might choose to fall back on this free digital education platform in these difficult times.  But the new research on the difficulty of protecting students’ data is not the first serious critique of Summit Learning.


A year ago, as part of an in-depth evaluation of the Providence, Rhode Island public schools, researchers from the Johns Hopkins Institute for Education Policy condemned the use of Summit Learning in that city’s public schools for academic reasons. In extensive classroom visits, researchers observed that students were neither engaged by their online lessons nor learning very much:  “(W)e witnessed significant problems in the use of the Summit Learning Platform. In one school, Summit was the major mode of mathematics instruction; in other classrooms, it seemed to be used for supplemental (e.g., remedial or practice) instruction. When we observed students using Summit, they were not engaged with the software in optimal ways. Instead of watching videos or reading tutorial texts, students went straight to the exam and attempted to answer the questions. When they answered incorrectly, corrective text popped up, which students did read; they then tried again with the next question. Even if students progressed according to plan, their learning would be limited to how to answer problems in the format presented by the Summit exam. In one school, we did not observe a single Summit math teacher engage in whole-class or even small-group math instruction. Instead, teachers either completed work at their desks, and/or answered questions when students raised their hand. Finally, the lack of teacher surveillance of student progress in some Summit classrooms meant that students worked very slowly through the material.”

It is, therefore not surprising that the Johns Hopkins researchers discovered students hated the program: “Students almost universally disliked the Summit program. They told the team that they are burned-out through the overuse of screen time, and bored. Some claimed that students actively left school as a result of the platform. There were classes we visited in which teachers appropriately integrated a blended learning model, but in most cases, students were just staring at the screens, totally disengaged.”

In 2018, students at New York’s Brooklyn Secondary School of Journalism walked out of school in protest and formally wrote to Mark Zuckerberg to critique the program: “Unfortunately we didn’t have a good experience using the program, which requires hours of classroom time sitting in front of computers. Not all students would receive computers, the assignments are boring, and it’s too easy to pass and even cheat on the assessments. Students feel as if they are not learning anything and that the program isn’t preparing them for the Regents exams they need to pass to graduate. Most importantly, the entire program eliminates much of the human interaction, teacher support, and discussion and debate with our peers that we need in order to improve our critical thinking.”  In their letter, the Brooklyn students also demanded information about the massive amount of student data being collected by Summit Learning.

A New York advocacy organization, the Parent Coalition for Student Privacy, has investigated and protested Summit Learning’s collection of student data: “Summit claims the right to collect an extraordinary amount of personal student information. Among the data collected by Summit… include: student and parent names and e-mail addresses; student ID numbers, attendance, suspension and expulsion records; their disabilities, gender, race, ethnicity and socioeconomic status; their date of birth, teacher observations, their communications with teachers; their grade promotion or retention, test scores, college admissions, survey responses, homework assignments, and extracurricular activities they participate in. Summit plans to track students even after they graduate from high schools, including their college attendance and eventual careers. Summit shares this data with as many as 20 corporate ‘partners’ including CZI (the Chan-Zuckerberg Initiative) and Google….”

NEPC’s New Report by Faith Boninger,  Alex Molnar, and Christopher M. Saldana

In their new report reflecting five years of research, Boninger, Molnar, and Saldana emphasize Summit Learning’s serious failure to protect student privacy: “Our review of summit partner school contracts suggests that student data collected by the Summit Learning Platform… presents a potentially significant risk to student privacy and opens the door to the exploitation of these data by the Chan Zuckerberg Initiative and possibly by unknown third parties—for purposes that have noting to do with improving the quality of those students’ educations.”

Summarizing the report in their guest column published by Valerie Strauss, Boninger and Molnar remind readers: “Mark Zuckerberg and Facebook have taught the world that data are fungible and can mean big money. And, also that data can be very dangerous when controlled by an opaque organization immune to public oversight. Regardless of who is named the owner of student data in partner school contracts, as Summit Learning’s technology partner, the Chan Zuckerberg Initiative has full access to the de-identified data and certainly has access to the technical expertise to re-identify it. Though it is often thought of as a charity, it is important to note that the Chan Zuckerberg Initiative is nether a charity nor a philanthropic organization. It is a limited liability corporation. A business.  As a result, although it may make charitable contributions, it may also make political contributions, engage in political lobbying, and invest in for-profit companies. Among the things the Chan Zuckerberg Initiative can gain from its collaboration with Summit Learning is access to significant amounts of student data that it can convert into a considerable amount of money.”

In their report, NEPC researchers recommend that state departments of education upgrade their oversight of personalized learning platforms “to require that the digital personalized learning programs be externally reviewed and approved by independent third-party education experts…; require that the assumptions and programming of all algorithms associated with personalized learning materials be audited for bias and other possible risks to students before the algorithms are implemented; and develop a standard data-security agreement that protects the privacy and limits the use of all data, including de-identified data, collected by schools through personalized learning materials….”  (emphasis in the original)

The researchers report that Summit Learning did not comply with their requests for data and detailed information. Data shared was anecdotal. Summit Learning persisted in undocumented claims that its learning platform is: “a ‘science-based’ personalized learning mode of teaching and learning that results in all of its students being academically prepared for college… None of the claims made by Summit Public Schools have been confirmed by independent evaluators.”

Boninger, Molnar, and Saldana conclude:  “The rapid spread of the Summit Learning Program—despite a lack of transparency and the absence of convincing evidence that it can deliver on its promises—provides a powerful example of how policymakers are challenged when faced with a well-financed and self-interested push for schools to adopt digital personalized learning programs.  There is now an urgent need for policymakers to move quickly to protect the public interest by establishing oversight….”