In Maryland, a Baltimore County student was handcuffed by police last year after an AI system misconstrued a Doritos bag as a weapon.
The calls for reconsidering how the student’s high school uses the technology were not new. Only two years ago, the federal government went after one company that sells artificial intelligence-powered systems to schools for misrepresenting how effective the tool was.
Before the introduction of this surveillance, it would have been difficult to imagine students being suspended for an unsent email, for instance.
But that is happening.
Outside of schools, surveillance tools have also supercharged immigration crackdowns, leading to the detention of young children of color such as Liam Conejo Ramos, a 5-year-old who was snatched by immigration enforcement while wearing a bunny hat.
From AI-powered monitoring tools and classroom cameras to digital tracking software and behavioral threat assessments, the increase in surveillance of American schoolchildren has been linked to higher levels of discipline and police interaction, especially for marginalized students.
It has also chilled student speech.
Schools use these tools in the name of safety and security.
Yet, civil rights advocates, educators, and researchers have raised concerns that these systems may reinforce existing disparities in discipline and contribute to the school-to-prison pipeline, where students of color are disproportionately pushed onto pathways that lead to the carceral system.
As the fears over student safety have grown, sales of surveillance technologies to schools have become immensely profitable. In 2021, for instance, the industry was estimated to be worth $3.1 billion, even as critics argued that it helped create a virtual panopticon, where students’ online activity is constantly watched.
Surveillance of all stripes is dramatically changing the educational experience for millions of students, especially those who are from communities of color, who are most likely to be monitored and punished. Some experts worry this is also leading young people to become increasingly inured to constant monitoring, eroding their First Amendment-protected freedom of speech as well as their ability to participate in American democracy.
For marginalized students, the tools may also exacerbate structural racism.
The growth of school surveillance is occurring when a climate of fear stemming from immigration raids has led to dramatic declines in school attendance. Immigration enforcement have accessed Flock Safety security cameras in schools to aid deportations amid President Donald Trump’s sweeping deportation efforts.
Some preliminary research suggests that these tools are making this worse. For instance, online surveillance tools may lead to an interaction with immigration enforcement: According to a 2025 report from The Center for Democracy and Technology, 7% of students and 6% of teachers report that a student was contacted by immigration enforcement based on what a monitoring tool noticed they were doing online.
It’s also occurring when sexual identity, or access to reproductive information, is being challenged in Republican-led states.
Monitoring companies argue that their products prevent LGBTQ+ youth from suicide. But these youth have said that these tools disclose their sexual orientation or gender identity without their permission, potentially placing them in a dangerous situation.
Lawmakers have charged schools to improve student safety amid pervasive school shootings and uneasiness over youth mental health. During the COVID-19 era, surveillance tools also helped schools supervise students when schools were suddenly forced to switch to remote learning.
Now that most students are no longer forced to learn from home, schools are relying on the technology as much as ever, notes Kristin Woelfel, policy counsel for the nonprofit Center for Democracy and Technology.
Companies like Omnilert and Evolv, which are tied to law enforcement, sell gun detection services. Companies like HALO Smart Sensor and FlySense sell vape detection services. And companies like GoGuardian sell monitoring services that keep tabs on students’ online activity.
The reach of these companies is sweeping. Securly, which sells a service that filters web content and monitors student activity, claims it has peered into more than 2.2 million interactions with AI on school-issued devices. Of these, the company estimates that 2% were flagged for “self-harm, violence or bullying.”
Federal data doesn’t meaningfully track the disparate impact of surveillance technologies in schools, according to Clarence Okoh, national director for civil rights and technology for TechTonic Justice, a community-based organizing effort.
What federal data does suggest is that surveillance of all kinds is pervasive, with an estimate from the National Center for Education Statistics finding that 86% of students between 12 and 18 years old had observed at least one security camera in their classroom in 2019. (Only a slightly smaller percentage, 75%, noted that their school had security guards or assigned police officers.)
Other data suggests that the connection between police and student monitoring is widespread. For example: Monitoring tools have also led to increased interactions with police more broadly, with 25% of teachers (and 18% of students) saying they know of at least one student who’s been contacted by the police or a school resource officer because of the student’s online activity, according to the 2025 Center for Democracy and Technology report.
Although they are marketed for safety, these tools have notified police before a student’s counselor or school when possible self-harm or violence was flagged.
In theory, more data could help schools to figure out what reforms they need to make so that they can better serve some marginalized students, says Jonathan Collins, executive director of the Center for Educational Equity at Columbia University Teachers College.
Tech platforms have sold their surveillance products to schools by arguing that they can reduce achievement gaps, improve test scores, improve school climate, or even save lives.
But so far, there’s been little evidence of the effectiveness of these tools, Collins says.
The NAACP Legal Defense and Education Fund has also recently argued that surveillance tools take money away from safety measures that have been proven effective, such as counselors, social workers, and restorative justice coordinators.
The scale of the data collection also raises questions about schools’ ability to manage data-sharing agreements with these companies to ensure that students’ information will not be sold, exposing them to exploitative consumerism, or to hackers when there’s a data breach, says Collins of Columbia University.
Okoh says that he’s seen lots of anecdotal accounts that algorithmic surveillance tends to get rolled out first in communities of color where the practices expect to get less pushback.
“Schools aren’t supposed to further the school-to-prison pipeline. They’re supposed to help students, to have the best interests of students at heart,” Amelia Vance, president of the Public Interest Privacy Center, previously told EdSurge: “And particularly when we’re talking about criminalizing students based on mental health issues, you get into some significant questions about whether schools are doing more harm than help.”
Amid big federal changes, some see this moment as “the perfect storm” for harming marginalized students.
The Trump administration has drastically weakened the federal agency that supervises education in the country, including radical alterations to the federal body handling students’ civil rights. It has also tried to undermine “disparate impact,” a major legal tool for identifying racism. Disparate impact uses the effect, rather than the intent, to judge whether a policy is racist. Some note this tool is crucial for fighting against algorithmic racism, since machines do not have “intent.”
At the same time, the administration has also pushed an aggressive AI agenda for schools, with financial support from major tech and education companies such as Google, Code.org, Microsoft and Amazon. That push has included deregulation of the sector and immense pressure for schools to adopt AI. This has even led to some pushback from teachers’ unions. One argued that the country failed to learn the lesson from social media, where a failure to regulate and a zeal for new technology led to a youth mental health crisis.
For those focused on what this means for marginalized students, it raises alarm bells.
Schools are being encouraged by policymakers to integrate AI across all of their systems, including school discipline, safety, and policing, says Okoh, of TechTonic Justice.
It means that schools have become testing grounds for some of the most legally suspect and ethically dubious technologies, he adds.
Technologies that algorithmically store and target students on the basis of protected characteristics end up disconnecting young people from the right to a public education, Okoh says. These students have to spend time outside of classrooms because an AI surveillance system flagged them as having made a threat when they were just making a joke on a school-issued device, he says.
With unprecedented access to what a student is doing and thinking, there are more chances to discipline students as well, argues Woelfel, of the Center for Democracy and Technology.
The pace of these new technologies also means that schools could cut humans out of the loop on things that are highly sensitive for some communities of color.
At the beginning of this year, Minneapolis Public Schools even solicited proposals to explore using AI to automate federally mandated accommodation plans for students with disabilities, which Okoh argues would have weakened their educational experience. Notable vendors in the education space, including Amplify Education, Kaplan Early Learning Company, McGraw Hill, and Savvas Learning, put in proposals, according to public records published on the website MuckRock.
And there are likely hundreds of school districts around the country considering this, Okoh estimates.
The big issue is that the country hasn’t fixed the biases that feed into the criminal legal system, says Collins of Columbia University.
But this is not a recent problem. Analysis going back years, relying on federal data, suggests that just having the police regularly at school leads to school officials handing over students to law enforcement even for minor infractions of school rules in a way that disproportionately harms youth of color. And data collected by companies like Apple, Google, and Facebook can contribute to over-policing communities of color.
These developments fuel worry that a womb-to-tomb approach to surveillance could also lead young people to become increasingly inured to constant monitoring and to the erosion of some of the basic freedoms, perhaps even their ability to participate in democracy.
For kids of color, every interaction with law enforcement increases the likelihood of incarceration, which significantly increases the chances that their voting rights will be restricted, Collins says.
What happens, he asks, when these kids exist in a digital ecosystem that mass accumulates data about them and which could politically disenfranchise them?
