Reports of AI-generated child sexual abuse material have skyrocketed in just two years

SAVANNAH, Ga. (WTOC) - Experts say a disturbing trend of the use of AI to create child sexual abuse material is becoming more common as technologies develop.

Zach Lowe, an Effingham County father, says he and his wife post pictures of their two daughters on social media to keep up with family and friends.

“Before, you just had to worry about, you know, is this picture appropriate to put on the internet? You know, if it’s a bathing suit picture or something like that, you know. Now, it doesn’t really matter. You know, you can take someone’s likeness and steal it, pretty much, and use it in any sort of way possible,” says Lowe.

Lowe says posting pictures of his daughters is now something he and his wife think twice about, and they’re even thinking about removing pictures they’ve already posted.

“It’s something we’re considering, and it’s something we will continue to consider and pray about. But if we deem it where the risk is not worth the connectivity, then we’ll do so,” Lowe says.

The data shows it’s a risk that’s greatly increasing.

The National Center for Missing and Exploited Children (NCMEC), which provides assistance to victims, families, and law enforcement, has a tipline for people to report when children are being exploited.

NCMEC first started receiving reports of AI being used to create child sexual abuse material (the expert-preferred term for child pornography) in 2023, receiving 4,700 reports. In 2024, that number growing to 67,000, and in the first half of 2025 it was already at 440,000 reports. That’s a 9,261% increase since 2023.

“There’s a variety of situations where individuals could be using generative AI to create child sexual abuse material or those guides to have those types of inappropriate conversations,” says John Shehan, a Senior Vice President of NCMEC.

Shehan explains that AI models can be used in several ways that are sexually exploitative to children- some, involving photos, and some involving commanding the model to interact with the user as if it were a child.

“There’s situations where individuals are texting to chat. So what I mean by that is they’re entering text into a chat model to pretend it is a child, trying to get that model to pretend is it a child, and to engage in sexually explicit chat. There’s other situations, let’s say text-to-text, where they’re entering text into these models, and they’re literally trying to generate guides and tutorials, trying to get suggestions perhaps on how to groom or sexually abuse, torture, even kill children,” says Shehan. “There’s some that you’re probably thinking of more commonly where it’s text-to-image. So individuals entering text prompts into these technologies in an attempt to get it to generate child sexual abuse imagery, or to alter previously uploaded files that they’re putting into it as well to make sexually explicit content.”

Shehan says for parents, the potential of their children’s photos being used illicitly is something to be aware of, especially with reports of these uses of AI exploding.

“It’s terrifying, honestly. It is quite terrifying. And I don’t know, you know, it’s kind of like medicine. You don’t know the reactions or the long-term effects until later on down the road. This isn’t something I don’t think we can afford to wait and see what the long-term effects are of somebody’s picture being used out of context or in a way that’s sick and vile,” says Lowe.

mAnIpulated: An InvestigateTV Series on the Impact of AI

Lowe says he doesn’t believe our laws have caught up to rapidly-progressing AI technologies, something Sheehan agrees with.

Shehan provided written testimony to the House Oversight and Accountability Committee earlier this year, recommending several protections against the use of AI to create child sexual abuse material (CSAM).

Some of the recommendations focused on the responsibility of AI companies to make sure their tech isn’t used in this way, even citing an example in which an AI learning model was trained using CSAM images.

“They’re the ones building these tools and technologies. They have a corporate responsibility to ensure that they’re not moving too fast. They’re taking the safeguards to ensure that tools like this cannot quickly and easily create child sexual abuse material, can’t be misused for these purposes,” says Shehan.

Shehan says parents should be vigilant, and understand the risks of posting pictures of their kids.

With even something as innocent as a picture of your kids on vacation becoming a risk, Lowe believes the problem is deeper than the deep web.

“I think we must look at the root as well, the root cause of this. You know, why have people gone out of their way to do things that are sick, like sexualizing children or meeting up with them online? I think it’s, at the core, it’s a heart issue,” says Lowe.

If your child has been a victim of real or fake sexual images being posted anywhere online, NCMEC has a program called Take It Down that’ll help scrub the internet of that picture.

There is a similar resource available for adults called Stop Non-Consensual Image Abuse.

If you know of any children being abused, you can report it to NCMEC by calling their hotline a 1-800-THE-LOST or using their cyber tipline.