Unlike anything the world has known before, AI has reinvented stranger danger. At the click of a button, anyone can make photographs or videos exploiting children. Once those graphics are created, they never leave the internet. The “deepfake apocalypse,” as Mara Wilson from The Guardian calls it, is a concern of parents around the globe.
The unknown content of AI is what scares parents. The future of their children is at stake, and with the wrong request of a website or app like ChatGPT, anyone’s future can be ruined.
In the Jewish world, the Masechet Derech Eretz—a minor tractate attached to the Talmud—teaches Jews to not be naive. In Chapter Five, the Derech Eretz Rabbah wrote, “כל בני אדם בעיניך כלסטים“ which translates to, “Let every person be in your eyes like a bandit / potential thief.” Though this was written over 1,000 years ago, this warning is more relevant than ever now with the acute dangers AI poses to children and the vigilance parents need. Parents cannot be completely trusting of the unknown capabilities of the internet and what it can do with their children’s images.
HBHA’s school nurse, Diane Svensson, a Johnson County mother of a seven- and nine-year-old, is extremely cautious with how her kids interact with social media. Svensson’s kids are “not at the point of quite knowing the extent of the internet.” However, she teaches her fourth grader “a lot about what we refer to as tricky people,” and the horrendous capabilities of strangers.
Svensson is not the only parent concerned about AI endangering kids. HBHA teacher, Micah Margolies, a parent of a two- and six-year-old, said his concerns about the dangers of AI’s effect on his kids are “similar to my concerns regarding everyone, but I think [they are] exacerbated and amplified with regard to kids.”
He discussed his concerns about AI modifying anyone’s images: AI can manipulate any image or video to show a person doing or saying something that never happened, or put a person in a place they have never been with people they have never met. Even the slightest tweaks AI makes to a photo can have a drastic effect on the meaning and change the context.
AI products and/or information rely solely on what their trained datasets contain. In 2023, a study at Stanford University found that popular AI training datasets contain over 1,000 cases of Child Sexual Abuse Material (CSAM), which is real or fake—including AI—evidence of child sexual abuse, according to Wilson. “It’s a crime to create, distribute, or possess” these images or videos according to RAINN.
Margolies also raised ethical concerns. He said, “A major fear of mine with AI stuff has to do with creating sexual imagery and videos.”
There are safeguards regarding sexual AI imagery, but on websites like ChatGPT, someone can ask the website to create sexual material with images of random people.
The new issue brought by AI is the fake imagery and videos, and how accessible it is for anyone to abuse another human or child with sexual images.
As parents, Svensson and her husband are cautious because they did not experience childhood with social media.
“In my heart of hearts, do I think that some person out there is gonna find a picture of my child and use it for harm? No,” but she will not risk anything for her children’s futures.
No one knows who could be dangerous. Face-to-face, 90 percent of child abuse victims knew their predator, according to the U.S. Department of Justice. Therefore, it is not surprising that in many deepfake AI cases, it is the victim’s peers who wrong them (Thorn).
Peers have quick access to someone’s original photos, making it easier for them to have the photos modified by AI. In a Thorn study, “one in ten minors said they personally know someone who has used AI tools to generate nude images of other kids.”
This reality comes with added worries and rules for parents. Svensson has more protection of what pictures and videos of her kids get posted. Svensson’s ‘parent anxiety brain’ makes her worry about the infinite what-ifs of what AI could do with pictures of her children. There are infinite photos and videos on the internet, and no parent wants their child to be the unlucky one who gets hurt. Svensson said she is “not willing to take that risk.”
Once someone makes an inappropriate image on AI, the picture is for the world to see. Sexual predators look for access, and in today’s day and age, access to real or fake nudes does not make a difference (Wilson).
When looking at this issue in a broader view, AI’s ability to morph any image can affect anyone’s digital footprint. This is something that, as Svensson mentioned, could be detrimental to someone’s career and life as a whole. Is there a way to regulate these datasets? How are parents supposed to protect their children from an online universe that has more power over their children’s lives than any living human holds?
