The “wall of silence” is crumbling. For the first time, Silicon Valley’s technology companies have begun releasing data about the diversity of their employees. The disclosures help to quantify a reality that many have suspected or experienced firsthand: Less than five percent of the collective workforces of Facebook, Google and Yahoo are black or Hispanic; four out of five team members are white or Asian men.

These statistics mirror broader trends. Only 13 percent of science, technology, engineering and math (STEM) jobs nationwide are held by black or Hispanic workers; at the same time skills related to STEM are “particularly scarce relative to [employer] demand.”

Gaps begin early in the education pipeline and persist into adulthood: Black and Hispanic students (who account for more than 40 percent of the K-12 population) are not only less likely to be prepared for Kindergarten, graduate from high school or attend “elite and competitive” colleges; they are also less likely to study subjects like computer science or to be part of successful startups. (According to a report by CB Insights, more than 80 percent of venture capital-backed early-stage companies have all-white teams.)

These disparities stem in part from phenomena such as implicit biases and cultures of insularity. They are also driven by (as well as reinforce) cycles of poverty. In the past 40 years, the proportion of students qualifying for free or reduced-price lunch has doubled to 70 percent, as racially-defined gaps in income and wealth continue to grow.

* * * * *

As important as these issues are in regard to consumer technologies, they are perhaps even more critical when considering the users and makers of edtech, an industry in which the holy grail of technology-enabled “personalization” risks becoming “an empty vessel into which one may pour any number of competing theories of learning or favored education policies.”

Increasingly, students’ academic and social experiences are influenced by edtech programmers, whose perspectives and priorities may differ significantly from users’. Nonetheless children from low-income households spend more time than their higher-income peers consuming media each day. As Hack Education’s Audrey Watters asks, “Who writes all these algorithms that will ‘personalize’ our learning through technology? […] For whom is ‘personalization’ defined (and by extension, for whom is ‘personalization’ programmed)?”

When it comes to edtech, Silicon Valley may be “leading the charge” in some ways (including job growth); nevertheless its contributions must be contextualized within “knowledge and a library of evidence far broader than that which the disruptive narrative includes.” If the “edtech revolution” is to fulfill its promise of closing gaps, it must be understood as part of a broader struggle toward equity and inclusion.

In order to analyze the social impact of edtech, we must begin by disentangling access and usage. Access is necessary, but not sufficient, for improving outcomes; variations in usage matter. Learning environments are characterized by “envelopes” which consist of both technology resources and social resources (e.g. scaffolding, modeling and support from parents and teachers) — these play a prominent role in shaping individual students’ educational possibilities.

Despite increased access to mobile devices and the rising (if tentative) popularity of BYOD, there are barriers that prevent many low-income children from accessing high-quality digital content at home and at school. And even where access exists, disparities can remain: Research has shown that gaps in resource allocation can exist even between students within a single school.

What can edtech developers do to encourage equitable usage of their products? One crucial step is establishing authentic communication with students, teachers and parents from low-income schools — this is the only way to ensure that real problems are being solved. Startups can also take steps to reduce friction to distribution and adoption in a wide variety of classrooms. In addition, early stage companies can prioritize the recruitment of culturally and socioeconomically diverse teams as well as investors, which may increasingly include non-profit foundations.

* * * * *

And yet, there are more systemic issues to address related to diversity in the design and usage of edtech. Enhanced empathy is important, but it only goes so far. In order to create outcomes that are truly equitable, edtech decision-makers must work to redefine the relationship between students — especially low-income students — and technology.

Educator Jose Vilson argues that “the lack of teachers of color is a symptom and not a cause of the education gaps we currently see.” In the case of edtech, it seems plausible that a lack of representation by (instructional and technological) designers of color is both a symptom and a cause of unfulfilled demand for content and products that reflect the diverse experiences and needs of low-income users.

There is work to be done combating conditions that inhibit equity and diversity at the professional level. In the meantime, a growing group of organizations is addressing these challenges through community-based, technology-enabled educational interventions. Black Girls Code teaches young girls and pre-teens of color in-demand skills in technology and computer programming. Science Genius leverages hip hop pedagogy to engage urban youth and educators in STEM exploration. Hack the Hood connects youth to real-world consulting projects building websites for local businesses and nonprofits. Qeyno Labs harnesses the interests of high potential youth from low-opportunity settings through radically inclusive hackathons. And new ideas are being generated all the time.

In a world where “genius is not always defined by academic success”, important questions remain about how to assess the educational value of technology. In some classrooms, consumer trends like user-generated content could intersect with ideas like portfolio-based assessment to create opportunities for more diverse interpretations of “achievement”. But to become genuinely transformative, educational technologies must function as part of a system of supports that equip low-income students with the skills they need to identify and solve challenges in their own homes and communities.

Cameron White is Associate Director of co.lab, an edtech accelerator by NewSchools Venture Fund and Zynga.org.

 * * * * *

If you’re interested in the roles that STEM and creativity can play in empowering youth, check out the video below from Summit 2014 panel “From STEM to STEAM: Creativity as a Lever for Social Justice.”