What happens when the era of AI-powered surveillance coincides with an authoritarian assault on public education? The transparency needed to answer that question does not exist, which is why the Knight First Amendment Institute, where I work, has filed a lawsuit seeking information about how one school district uses surveillance systems to monitor student laptops.

It is estimated that millions of children—nearly half of K-12 students across the nation, according to a recent New York Times report—are subject to digital surveillance systems that can potentially monitor every word or phrase they type on school-issued laptops, tablets, and software. These software systems, supplied to school districts by private education technology companies like GoGuardian, Gaggle, and Lightspeed, scan students’ communications, internet searches, and assignments, searching for keywords or phrases that may indicate cyberbullying, thoughts of self-harm, or thoughts of harming others. If a student uses a certain word or phrase deemed inappropriate by the vendor or the school district using the technology, administrators are notified and, in some instances, police get involved. The systems also have filtering capabilities, which can restrict students’ access to certain websites and pages based on their content.

Given the widespread concerns about youth mental health, many people might view such digital surveillance as a godsend, a critical tool for combatting tragically high rates of youth suicide and depression, as well as school shootings. But that viewpoint likely rests on two assumptions: first, that the systems are as effective as the edtech industry claims, and second, that school districts are limiting the systems’ purview to content that relates to student safety.

It's not clear, however, that those assumptions have merit. The scant information we have about school surveillance suggests that the systems can lead to erroneous flags and can be used as a tool of overzealous surveillance by school administrators. Public records obtained by the Electronic Frontier Foundation have exposed several examples of students being flagged for innocuous visits to websites containing the text of Genesis from Bible.com, classic literature like Romeo and Juliet, and publications about Martin Luther King Jr. and the Civil Rights Movement. In other cases, surveillance of students’ online activity has led to the unwanted disclosure of private details about their sexuality and the flagging or blocking of race-related content, according to a report from the Center for Democracy & Technology. And just last year, journalism students at a Lawrence, Kansas, high school successfully campaigned to get their files omitted from the purview of Gaggle, arguing that the school’s use of the technology would have a chilling effect on critical reporting on district staff and administration.

Despite the edtech industry’s claims that their artificial intelligence products have saved thousands of students’ lives, the Associated Press reports that no independent research has corroborated their efficacy. Industry data regarding the frequency and accuracy of the systems’ alerts is usually kept hidden behind closed doors, exclusively in the hands of the for-profit companies that develop and market them. And while school districts maintain that their use of these tools is aimed at deterring self-harm, cyberbullying, and other harmful situations, their ability to customize the standard list of keywords and blocked websites provided by edtech companies raises serious concerns about the extent of surveillance and censorship school administrators can carry out, and the potential impact on students’ privacy, speech, and associational rights. For example, school districts could easily block “Black Lives Matter” websites, as at least one district reportedly did in the past, and flag any research or discussion of sexual orientation. In a moment where many children are prohibited from discussing racial history, learning about gender identity, or reading banned books in the classroom, it is critical that they have alternative channels for exploration and expression, including the internet and messaging platforms. Digital surveillance endangers both the sanctity of their private communications and their freedom to access more accurate, complex, and engaging ideas than their schools might permit.

These concerns have intensified in recent weeks as the Trump administration escalates efforts to control how K-12 schools teach history, race, and gender. It’s not clear what role, if any, schools’ deeply entrenched surveillance capabilities will play in their efforts to comply, but it’s important that the public knows about it.

In an effort to learn more about how schools are monitoring our students, The Knight Institute submitted a records request to the Grapevine Colleyville Independent School District in Texas. The district has resisted our efforts to obtain information about student surveillance, claiming the records are statutorily exempt from disclosure. The district’s position is that all information about the keywords and websites it monitors or blocks falls within the public records statute's computer network security exception, and disclosure of that information would expose the district’s network to security threats. But it’s not clear how transparency about the scope of student surveillance would lead to a security incident, and the district has offered no basis for its contention that it could. Until the public has access to this information, the extent of the district’s monitoring remains unknown.

We’re hopeful that our new lawsuit succeeds in its efforts to compel the disclosure of this information at a critically important time. The public—particularly the students, parents, and teachers most directly impacted by this technology—deserves to know whether tools that purport to protect student safety are being exploited to stifle and censor student expression.

Editor's note: Gaggle, GoGuardian, and the Grapevine Colleyville Independent School District did not respond to Teen Vogue's request for comment. In an email, Lightspeed chief of staff Amy Bennett said, “Monitoring activity on district devices to keep students safe isn’t just the right thing to do—it’s also required by federal law under the Children’s Internet Protection Act (CIPA). To minimize false alerts, Lightspeed Alert uses a combination of AI monitoring and trained human Safety Specialists.” Bennett added that their programs are “built with Privacy by Design. We hide personally identifiable information (PII), share only the minimum data needed with only the personnel required, and fiercely secure our systems. And our AI is specifically trained to look for signs of distress and potential harm—not things like bad words or flirting.”