The Political Economy of Education Technology
The trend towards a greater presence of digital platforms in higher education has accelerated during the COVID-19 pandemic. Despite ongoing growing pains, remote learning tools have enabled many institutions to remain operational. The more or less seamless transition to virtual classrooms is largely owed to the ubiquitous use of Learning Management Systems (LMS), which have evolved from basic teaching support to robust infrastructures. According to estimates, already before the pandemic 99% of colleges and universities had integrated school-wide LMS while 85% of instructors had used LMS in class, whether in-person or distance learning, synchronous or asynchronous content. By the end of 2019, Instructure Canvas alone registered nearly 20 million enrollees in colleges and universities. As the LMS market saturates and new LMS adoptions decline, companies compete by promising advances in the areas of interoperability, customization, analytics, and design. The pandemic emergency clearly illustrated, for instance, the importance of seamless integration between video conferencing platforms and LMS. But instead of simply enabling remote learning, new education technology (EdTech) reimagines the pedagogic environment, extends performance metrics into virtual classrooms, and reshapes modes of participation and academic labor.
For an already growing EdTech industry the pandemic turned out to be an exceptional boon. Against the backdrop of an estimated $76.4 billion market size in 2019, that is before the pandemic, analysts projected “a compound annual growth rate (CAGR) of 18.1% from 2020 to 2027”. But already the first three months in 2020 account for $3 billion of “global EdTech Venture Capital, nearly 10% of the prior decade’s total.” Meanwhile, as Ben Williamson notes, the proliferation of financial tools, EdTech exchange trading funds (ETFs), and “power networks” are already affecting education policy worldwide. It’s possible that market patterns, industry growth, and overall excitement about EdTech will undergo corrections once in-person classes resume, but many of the current developments will have lasting impacts.
Video Conferencing Tools: Panopto, Class, and Engageli
The video platform Zoom emerged as one of the decisive winners of widespread lockdown and shutdown policies in and beyond academia. Seemingly overnight, the company became the go-to solution for synchronous remote learning and reported $672 million in profits at the end of 2020, a 3,200% jump from the previous fiscal year. Zoom is part of a growing information ecosystem that includes video conferencing tools, asynchronous content creation, and course organization. In this space, several EdTech businesses now seek to capitalize on Zoom’s success or outright replace the platform.
Founded in 2007, Panopto provides a platform for capturing lectures, meetings, and events and managing recordings in a cloud archive. Among other functions, Panopto creates a searchable library that turns recordings into “important digital assets” including closed captioning and audio transcripts. Zoom only recently caught up in offering closed captions for free accounts. Lecture recordings, closed captions, and transcripts serve important goals of making education more accessible and inclusive across geographic regions, individual schedules, and different learning abilities. But as educational video content creation and management become more common, important questions remain: What are the mechanisms of informed consent among students and faculty when lectures are uploaded to cloud servers? How might pedagogic approaches change if recording and captioning become the norm? Who ultimately retains ownership of recorded materials that could serve as training data for proprietary machine learning applications?
Increased reliance on Zoom resulted in concerns about cybersecurity and emotional fatigue. The challenges of Zoom university led to a wave of add-ons and new platforms that promise to improve the Zoom experience. Class, a startup of Blackboard co-founder Michael Chasen, raised $58 million from Zoom board members, Salesforce Ventures, and Tom Brady. The certified Zoom reseller reorganizes the virtual classroom by enabling a view that more clearly differentiates between instructors, students, and teaching assistants. In addition, Class introduces attendance, participation, and engagement metrics into live meetings. Several of these features address known problems especially for instructors. Zoom’s gallery view, for instance, can have a flattening effect that hides instructors among students. Meanwhile, features that monitor student engagement seem more insidious. Tracking how long students speak in class and whether Zoom is the primary app on their computer might trigger anxieties about privacy and questionable performance proxies. Perhaps more importantly, Class extends the formal logic of LMS to video conferencing, promising more efficiency by automating the virtual classroom. However, EdTech’s promises of more efficiency and automation often translate to changing and/or additional work for educators: “automation in the pedagogic environment does not eliminate teachers’ labour, but reconfigures it by […] shifting their efforts from actual teaching to the 24/7 coordination, moderation and facilitation of student engagement.” As LMS and video conferencing continue to merge, so too will administrative activities shape pedagogic experiences.
Similar to Class, Engageli is a startup combining EdTech experience and venture capital. Unlike Class, Engageli does not integrate with Zoom, but seeks to replace it through a more intuitive and collaborative experience. The browser-based platform reimagines the classroom around tables where students might communicate while listening to lectures. Compared to Class and other video conferencing tools, Engageli de-emphasizes the live experience and favors a multimodal classroom environment in which students collaborate on Google Docs, annotate presentation slides, and discuss live or recorded lectures. Whether small seminar or large lecture, Engageli involves a higher learning curve that requires instructors to prefigure table arrangements and enable certain activities through integration. In other words, Engageli might entail similar preparation as creating an LMS course shell. As a hybrid of LMS and virtual learning software, Engageli at once drives and benefits from modularity whereby educational programs are broken down into smaller units that are, in turn, more easily integrated and automated.
Lifelong (and Modular) Learning
While developing professional skills for life remains the essential pitch of higher education, the corporate world in turn is discovering the idea of lifelong learning. Aside from equipping K-12 schools and colleges with LMS, Canvas now promises “big wins with Canvas corporate” and urges employers to “make learning an inherent part of your organization.” According to a recent white paper by the Longevity Project, the MOOC platform Coursera reported that 2020 enrollment “jumped 640% higher in March and April compared to the same period in 2019, growing from 1.6 to 10.3 million users.” Like other consultancies, the Longevity Project identifies lifelong learning as a practical response to a trend of accelerating automation and job insecurities. It seems like continuous education is the future of work, at least according to business leaders, consultancies, and sector agencies invested in tech-driven solutionism. But framing automation as an inevitable economic force that demands innovative responses not only limits decision-making to software providers and university leadership, but also glosses over how automation happens. As a growing body of research shows, automation routinely involves the work of integrating and maintaining new systems, which is often handed down the ranks. In other words, the grand visions of smart workplaces and smart campuses are predicated on the fragmentation of jobs into tasks, the deskilling of professions, and the large-scale production of data by increasingly invisibilized workforces. While a reasonable approach on the surface, lifelong learning is also a self fulfilling prophecy that legitimates more digital tools and more automation.
As education is framed as a continuous, lifelong effort, its labor systems seem increasingly time consuming. Studies suggest that pandemic-induced remote work has upped workload and stress of faculty, especially pre-tenure women. Of course, injunctions to work longer hours predate the pandemic and will surely outlast it. EdTech facilitates these trends, for instance, by extending LMS access on-the-go. Advertising its mobile app, Canvas seeks to “break down learning into smaller pieces (aka micro learning).” Such modularity or “content chunking” is a condition for both on-demand education and lifelong learning.
Student Privacy/Ethics
The growing boom of educational technologies brings with it an array of data ethics and privacy concerns, many of which deal with personalized learning and the collection of student data. Today, nearly all students in the U.S. are currently in mostly- or exclusively-online-only courses, and 85% of students indicate that they felt “the same or better” about their experience in online versus in-person instruction. This relays important beliefs or assumptions our students might have about being online; especially for today’s youngest university’s attendees, a version of the Internet without Big Data, algorithms, and intrusive surveillance seems impossible to imagine, let alone build. While students are broadly aware of such issues, they are often wholeheartedly surprised to learn just how granularly their online activities are followed, even while writing an essay or completing a quiz–and we cannot expect students to make informed decisions about Online Learning when the data-related truths of these spaces are not made clear.
Due in part to the ever-increasing pressure to keep enrollment high, students’ socio-emotional “experience” is viewed as an especially attractive measure for colleges to track overall “performance” and predict what might bring students in future years. Because of the pandemic, low-cost and close-to-home options matter more to students and parents alike than on-campus attractions at far-away schools; a university that might have, pre-COVID, advertised rock climbing walls or on-campus fine dining to attract students from all over the world is highly incentivized to track and measure those same students’ “satisfaction” with the school now that these features are no longer available.
Data relating to student satisfaction and “experience” within university settings is thus among the most lucrative and controversial applications of EdTech. In April 2020, just a month after most U.S. institutions shifted to exclusively or primarily Online Learning, 92% of colleges and universities marked “students’ mental health” as a top immediate concern, and 88% marked “employees’ mental health.” Students’ data is acquired via a Learning or Experience Management Software about topics ranging from their feelings, demographics, academic performance, living situations, and overall “experience” at school or with coursework. This happens for individual course offerings at the end of a semester, when students are asked to quantify experiences relating to what they learned, how much time they spent on homework, and if they felt their instructor was “available” to them. Surveys are also sent to entire student bodies about myriad topics–their satisfaction with first-year orientation, their likelihood to attend summer courses, etc.
As a student, I have regularly been asked by Experience Management Software like Blue and Qualtrics about the quality of on-campus activities or how often I use the library–but I have also been asked if I am homeless or feel “unsafe” at home, if I am in an abusive relationship, if I struggle to feed myself, if I take any medications, how I identify my gender and sexuality, and indeed if I feel emotionally or affectively “satisfied” by my various relationships to my school. Surveys and responses just like this are already being used by for-profit organizations and schools all over the nation to, for example, rank universities on given criteria (i.e. US News & World Report’s list of “Colleges with Great First-Year Experiences”), fully redesign courses where students usually test poorly, or even predictively flag individual students whose online “behavior” is associated with “low performance.” Here, the problem is not that colleges are trying to improve their facilities and course offerings–the problem is the profound lack of transparency and contextualization between for-profit software companies and ranking organizations, university administration, faculty, and students. Additionally, surveys are very far from the only collection going on; when a student agrees– as is made seemingly-compulsory by their school– to an LMS’s terms and agreements, they are often unwittingly agreeing to also being surveilled indefinitely and without notice, for aims potentially incongruous with or violent to their own, by varyingly opaque corporate entities.
These Learning Management Systems track potentially every move a student makes: their demographics and course history–but also how long they spend on each test question; if, when, and for how long they left the course page and opened a new browser window; if they message another student, and the contents of that message; a detailed history of their essay drafting process; if they refresh or reload a page. And that data does not simply disappear once it is used for the university’s administrative concerns—data collected about a student’s attendance or grades, and then sold to potential employers, might prevent them from getting a job years down the line. The troubling trend is toward monetizing and evaluating student data, wherein a student is measured against a set of evaluative criteria to determine their fit for a class, school, or any institutional setting. This is not unlike the creation of a credit score. And while a student is unlikely to have consented to this collection, they are likely beholden to whatever decisions are made as a consequence; universities increasingly turn to softwares’ predictive modeling for choices like determining a student’s financial aid package or preemptively flagging students who seem like they might drop a course midway through the semester.
Studies show, again and again, that data collected in order to rank and hierarchize is neither capacious nor reliable enough to inform major policy decisions–and yet, nearly every university in the nation is engaged in just this. Unpacking this complex issue requires a serious look at the flaws latent in how data about students is collected in the first place. We know from our Seminar guests that algorithms trained by data and deployed by software companies reproduce the very same biases and inequities as the world they exist within, and EdTech companies are certainly not excluded from this fact. Algorithms developed by EdTech companies are often trained using data from schools with very few students of color or low-income students, meaning that the algorithm is far less accurate and far more biased when it encounters such students. Furthermore, these softwares simply do not and cannot take into account the veritable world of variants and contexts that might be informing how any student is behaving, performing, or responding. So how are we to know if Student Y is genuinely displaying behaviors associated with “low performance,” or if Student Y happens to be having a really crummy week, or if Student Y is from a racial or socio-economic group that a software is biased against? The software cannot sort through these contexts on its own.
Impact on and Surveillance of Instructors
The future of EdTech also has serious implications for faculty-institution relationships, since instructors are rarely consulted prior to their employer’s adoption of an integrated EdTech software. So while Zoom might work best for one instructor’s setup, and Microsoft Teams might have the best features for another’s, instructors are often mandated to use whatever software their institution has already bought, regardless of fit. And so, an instructor’s role becomes not only to teach and assess, but to perform administrative tasks related to the management of a software that they may not be very interested in implementing in the first place. While studies report that teachers are significantly more confident in Online Learning environments than they were before COVID, only 26% of respondents to an August 2020 survey report using data to inform future course decisions–and a staggering 66% of instructors report having serious concerns about equity gaps between student groups as a result of Online Learning. Additionally, over half of instructors felt they had adequate training during COVID from their institution–but of course, this means that nearly half of instructors still do not feel adequately prepared, even as they are already engaged in online teaching.
A striking trend in these reports is that while instructors feel confident and, generally, well-trained in their use of EdTech tools and softwares, they do not actually feel such softwares are necessary or beneficial to their teaching–and, at the same time, voice serious concerns, over and over, for students who don’t have regular internet access, reliable computers or devices, or for whom distance learning is not equitable due to a wide variety of disabilities, socioeconomic statuses, and living arrangements. Much has been written over the years about how the U.S. does not listen to its teachers, both in K-12 and Higher Ed environments; high school teachers and university professors alike have seen a steady erosion of their workplace autonomy, benefits packages, security, and overall quality of life. A 2018 Inside Higher Ed study confirmed that over 70% of all faculty positions in higher education are off the tenure track, and half of those jobs are part-time only. And as COVID creates potentially years-long budget deficits, particularly for large four-year public schools, the numbers are likely to get even more bleak: adjunct faculty working conditions are historically terrible–and they’re far more likely to both teach online and receive inadequate training to do so.
In the face of these historic and rapid changes in the higher education arena, Universities–often under unrelenting pressure to fill classes and increase enrollment–turn to EdTech software to monitor and assess instructor performance. This happens not only by surveying students, but also directly surveilling instructors in their online environments. Universities ultimately have access to the same sort of information about what instructors do while logged into an EdTech software as they do about students: how many hours they’ve clocked grading or building lessons Canvas or Blackboard that week, the contents of their Zoom chats, any meetings they’ve scheduled through a University-affiliated booking system (like Outlook or Google Calendars), and potentially much more. And productivity demands for higher ed faculty are indeed higher than ever; older, tenured academics have even proclaimed that they would not have been able to get the job they have today if the standards of productivity had been the same back when they first got it. But while the bar for being productive “enough” keeps getting higher, faculty are also up against the factor of automation.
The more digital our education becomes, a range of daily assessments including anything from face-to-face conversations with students, quizzes, exams, exit tickets, etc.–ways that teachers learn to assess students that include but certainly cannot be limited to exam scenarios– are digitally transformed to simple online quizzes and tests. In these online environments, AI can be trained to monitor students’ eye movements, audio levels, internet activity, or other “warning signs” of cheating; this AI can cheaply and easily replace a human instructor who the university no longer needs to find, pay, or train over longer durations of time.
So who is EdTech for? Students report feeling utterly isolated, depleted, and hopeless in the online-only world of COVID higher ed, and the facts were not much better prior to it, either. As EdTech becomes a tool of productivity enforcement, punishment, and surveillance–rather than care–students have all-time-high rates of anxiety and depression as directly related to the “tremendous pressure” to perform well academically. And EdTech is certainly not made for teachers either. In fact, EdTech is typically developed straight out of software made for surveilling and policing employees at massive corporate entities. And if EdTech is neither for students nor teachers, the future we’re building together will entail an absolute paradigm shift in the field of education–one that prioritizes private software company growth over all else.
Conclusion
The quickly-growing body of critical research on EdTech will continue to illuminate the existing absences and issues in higher education, such as the hidden labor of instructors and spurious assumptions behind excessive data-driven predictions. Despite the pressing challenges of teaching and learning within the corporate-dominated EdTech sphere, students and faculty can still work together towards (cultures of) autonomy, serendipity, and humanity in the classroom. Even in 2021, it is not a given that Blackboard, Zoom, or Canvas (must) dominate our every educational moment, ; the production of resistance against, analysis of, and community-building in the face of EdTech made by educators and students all over the nation is a testament to exactly this. Narratives that “digital natives”– young people born after the Internet, smart phones, and social media–“don’t really care” about classroom surveillance or the future of online education, are evidently misguided.. Students are ready to talk about the changing face of education in the U.S., and so are educators. And the more productive conversations we initiate–between ourselves, our students, our institutions and our governments–the closer we can get to an equitable, safe use of EdTech.
This blog was co-authored by Mario Khreiche and Jane Rohrer. More about their work can be found here: