The debate about whether or not to post your kids’ faces on social media isn’t new. I have friends whose kids are in college now who I remember struggled with the decision about whether to post baby photos to Instagram.
But very recently, that conversation has taken a significant turn because of concerns about AI — and even prompted Meta to take action.
Two viral things have happened on Instagram in the last week or so, with pretty different origins but ultimately the same result: parents saying that they shouldn’t post images of their children online anymore.
The first was a viral Instagram from Kalie Robins, a mom in Utah. Robins was shocked and creeped out when she noticed Meta AI suggestions on a video she had posted of herself and her kid.
The Meta AI chatbot suggested questions like “who’s the child passenger?” Following more of the Meta AI questions, Robins found that it pulled photos of her that she believed had been deleted, and information about her kids from her mother’s Facebook page.
While yes, Meta AI was pulling information that already exists online, it was also suggesting questions about the identity of children — potentially suggesting that to strangers (it’s unclear if only Kalie saw this set of Meta AI questions or if a stranger would have been suggested different ones).
Meta told me that these AI suggestions were a mistake that had since been fixed.
“When someone clicks on a prompt and uses this feature, the response it generates is based on information the user already has access to,” said Meta spokesperson Andy Stone. “But we missed the mark because the feature never should have prompted the individual with questions like that and we’ve fixed that issue.”
Meta AI suggested questions about her kids
I spoke with Robins, and she expressed that she wasn’t unaware of how Meta AI was able to pull up the information. To her, it was a stark reminder of just how much of our data is on social media, even if we don’t realize it.
“I thought I was so careful because I never did the back-to-school photos with the school and teacher name. I never posted in front of any landmarks or in front of our house with street signs. I never did any of that,” she told me. “I thought I was being so careful and ultimately that didn’t matter. It didn’t even get that info from me. It got it from other people, so it didn’t matter how careful I was being in this case.”
The online reaction to Robins’ video was twofold: first, that Meta AI did something creepy and off-putting (which even Meta agrees with), and that this is a good reminder for parents to be wary of what information is out there about their kids and any photos they post.
Professional photographers are shying away from sharing kid photos
At the same time, another viral trend on Instagram is making people rethink how kids’ images are posted online.
A bunch of professional family photographers, nearly all women or mothers themselves, have been posting notices announcing a new policy that they won’t share kids’ faces on social media or their professional websites anymore.
The concern here is over AI-generated child sexual abuse material (CSAM), and the risk that photos posted to social media could be used to either create deepfakes of specific children, or that their faces could end up in data training for AI models that end up creating this kind of content.
This is a legitimate concern, and there have been criminal cases involving AI-generated CSAM, some gaining national media attention.
A recent legal ruling about AI CSAM
A judge’s recent ruling in a criminal case about AI-generated CSAM has gone viral in some Instagram parenting circles. The details of the legal ruling are complex (there’s some great detail here in the Washington Post story about it), but the short version is that the judge wrote that there’s a legal precedent that possession (not production or distribution) of AI-generated child pornography is protected as free speech.
The implications of the ruling — and what it means for innocent photos of kids — is what is now going viral on Instagram.
Hannah Haggard, a family and wedding photographer based in Kentucky, posted to her Instagram that she was changing her policies, and will not post any kids’ faces — even with parents’ permission.
“I’d been toying with the idea already,” Haggard told me. “I never post without consent, but I still was already considering it. Then once the ruling happened, it was confirmation that that’s what needed to happen.”
A lot of the online reaction I’ve seen seems outraged at the ruling, which at a glance seems to be good news for child predators. But Mariefaye Bechrakis, a lawyer specializing in human rights and child rights who has written about AI-generated CSAM legal issues for the American Bar Association, says that it’s more nuanced.
“What is important to note is that the judges are flashing a green light for a future appeal,” Bechrakis told me about her opinion of the judge’s ruling. “They basically gave a future roadmap and a real opportunity to future prosecutors. They noted that modern AI models don’t create images out of thin air, but that they must be trained on large datasets.”
Katey Sutton, a family photographer in Missouri, also posted about her policy of not showing kids’ faces on the internet.
“I don’t believe AI is truly ‘synthetic’,” she told me. “It’s known that AI scrapes the internet and uses information it’s given/finds to inform, recommend, create, etc…”
Take a scroll through Facebook or Instagram, and you’ll see that clearly not all parents have ceased posting kid pics. But for some people, these recent AI scares are making them think carefully about the risks before posting. And that’s probably a good thing.
Read the full article here
