Table of Contents
School Social Media in the Age of AI
We have reached the point in our rapidly-accelerating technological era when it is fair to ask, "Is it ethical for schools to post photos of students (or staff) on publicly-accessible social media?" Increasingly, the answer is no.
The question is raised by the recent arrest of Adan Covarrubias, 31, in Carrizo Springs, Texas. He was apprehended by agents with the Criminal Investigations Division of the Texas Department of Public Safety, with assistance from the Federal Bureau of Investigation and the Dimmit County Sheriff's Office.
Following a months-long investigation into AI-generated child pornography, a warrant was issued authorizing the FBI in Del Rio, Texas, to search Covarrubias's home.
Investigators said they identified more than 900 images and videos containing AI-generated pornographic images created by Covarrubias, depicting both prepubescent and pubescent children. DPS said multiple non-AI-generated pornographic images of children were also identified in the seized digital evidence.
The most disturbing part of the DPS report was the fact that Covarrubias had allegedly downloaded photos of students from school websites in the area and then used AI-powered "nudify" apps to generate child pornography images. Investigators have identified 30 victims so far.
"Demented Arts and Crafts Project"
The manipulation of non-explicit child photos for sexual gratification is not new. In 2013, for instance, a Spanish teacher in Chambersburg, Pennsylvania, was arrested on child pornography charges; police later determined that he had photoshopped the heads of 136 students onto images of child pornography. In 2020, an English sports coach was jailed for eight months for using Photoshop to alter photos of schoolgirls so that they appeared to be having sex with him. And in 2022, a teacher in Suffolk, England, was jailed for five years after a court found that he had taken a girl's photo from her school Remembrance Day event and used her face on several extremely graphic images.
These are not, unfortunately, anomalies. Numerous educators were arrested and convicted for manipulating images in the pre-AI era. The problem for schools is that we are long past the scissors-and-glue-stick, or even the Photoshop, phase of criminal image manipulation. Dozens of educators around the country (and in other nations as well) have been arrested over the last two years for using artificial intelligence to manipulate otherwise innocent images that schools routinely make public.
Schools must confront the fact that so-called artificial intelligence has brought a genuinely unprecedented level of automation to this crime.
Just under a year ago, the Center for Democracy & Technology surveyed the sharing of non-consensual intimate imagery (including morphed images and deepfakes) in K-12 schools. They concluded that "concerns over the widespread nature of NCII, both authentic and deepfake, in public K-12 schools across the country are well-founded." They also found that:
Schools are not doing enough to prevent students from sharing NCII: Very few teachers report that their schools have policies and procedures that proactively address the spread of authentic and deepfake NCII. Instead, schools reactively respond once there has been an incident at their school. This unfortunately leaves many students and parents in the dark and seeking answers from schools that are ill-equipped to provide them.
Stop Feeding the Beast
Share Photos Privately
The first and most important step is to minimize the number of photos that your school distributes publicly. There are a number of edtech vendors who offer schools the ability to communicate privately and securely with parents, rather than simply posting on social media sites like Facebook and Instagram.
Monitor and Audit School Social Media Accounts
Social media channels can play an important role in promoting school activities, providing information, and boosting community spirit. There's no reason to discontinue using social media for those purposes.
Schools should, however, stop distributing identifiable photos of students and staff on public-facing pages. They should also consider reviewing past social media posts and removing any photos that might potentially be misused. That is obviously a time-consuming process, and potentially a futile one, given the fact that posted photos may already have been downloaded.
Update Media Consent Language
There is a non-zero chance that at some point, a school district will face a lawsuit from a deepfake victim who alleges that the school was negligent in either posting her photo to social media, or failing to remove it as the scope of this threat became clear.
The first thing a school defense lawyer will do is look to see whether the staff member or student's parents signed a consent form allowing the school to photograph or take video of the individual in question.
Even if there is a scrawled signature on the form, however, that's not the end of the inquiry. Has the language of your release form been updated to cover the use of identifiable images on social media, where they may be permanently searchable and repeatedly downloaded? As we become better educated regarding the risks of image manipulation, parents, staff, and students will expect (and deserve) greater specificity about what can be posted and where.
Educate and Protect Staff as Well
While students tend to be the most common victims of deepfake manipulations, female staff are also at risk. The first responsibility of a school administration is to educate staff members about this potential problem. That has the collateral benefit of reminding potential perpetrators that administrators are on the lookout for misconduct. It also will encourage educators to rethink the images that they share on their personal social media accounts.
The second is to consider the same limitations on public social media posts that should be in place for students. There are many ways to illustrate classroom activities and school activities without creating risks for individuals: Wide-angle shots, backs-to-camera, hands-and-projects – all minimize the potential for misuse.
Pressure Vendors to Protect Images and Personal Data
Do schools have the ability to force companies to better protect identifiable student information, including images? That depends. Realistically, the ability of individual school districts to negotiate with a social media company like Meta (which owns both Facebook and Instagram) is basically non-existent. You may not like how they operate or their willingness to tolerate misuse of content, but there is not much you can do about it.
However, if a school district purchases edtech software designed to facilitate the confidential sharing of messages and media, then it has more leverage to negotiate how the company handles images and data. Districts should work with their attorneys to make sure that contracts contain language that holds vendors liable for failure to securely protect sensitive information.
Be Realistic about Limits of Prevention
The grim reality is that it is impossible for schools (or parents) to completely prevent the misuse of digital images of students. Using a "secure" communication system will cut down on the potential misuse of images, but not eliminate it entirely. Any time a digital image is created and shared, even privately, it can be copied, reshared, and then potentially misused. And this approach, unfortunately, does not eliminate the risk of misuse by educators and staff, who may have broader access than members of the public.
In some of the cases I've studied, the images were taken by spectators at sporting events, plays, or other school events open to the public. Are we moving towards a time when schools will limit attendance at events to verified family and friends?
And of course, there is little that schools can do directly about the images that students post to their personal social media accounts. In New Orleans, for instance, an educator named Benoit Cransac is currently being held on $8 million bond for creating deepfakes of female students using images that they posted to social media. His case is a disturbing window into the ready availability of digital content and the tools that can be used to manipulate it.
Schools cannot address all of these issues, of course; parents and students must reflect on their social media use as well. But in this moment, unfortunately, it is increasingly clear that schools do have an ethical responsibility to rethink their social media practices, and to take steps to minimize the risk of lasting harm to their students and staff.
