Discord is moving to a “teen-by-default” model worldwide. Starting with a phased global rollout in early March 2026, Discord will apply stricter default safety settings to new and existing accounts and will require age assurance to access 18+ channels, unblur flagged content, or to turn off certain safety protections.
This update is part of Discord’s global effort to comply with online safety laws and age-appropriate design standards. For parents, the practical reality is straightforward: teens do not need to “unlock” adult settings, so Discord’s default safety settings are worth keeping.
Discord says age assurance prompts can appear when an unverified user tries to:
If they do not verify as an adult, the usual outcome is simple: the protections stay on and adult-only features stay locked.
Related: What Parents Should Know About Discord Servers for Teens
Discord describes three methods:
Discord says the experience is designed so most users can keep using Discord without verifying; the practical consequence of refusing is that the “teen-by-default” protections stay in place and adult-only features remain inaccessible.
If age assurance fails, Discord indicates the user can retry (often with ID scan as a fallback), and if a user is verified as below the minimum required age in their country, the account will be banned with an appeal path.
At account creation, Discord requires a birthday and states that in some cases it may require additional information to verify age.
What to teach your kids before March:
Actions you can take:
And, finally, monitor their activity across Discord and all the other apps they use. BrightCanary is the most robust way to supervise your child’s activity on iOS, from keyboard monitoring to text message monitoring and emotional insights. Get started today with a free trial on the App Store.
Imagine this: you are scrolling on your phone after dinner when a parent shares a photo in the class group chat. The image shows your child behind the cafeteria, holding a vape, and your name is getting tagged in the conversation. Your child is right there in the doorway, insisting the photo is fake. You can feel the urge to respond immediately to it, ground your child, or take some kind of action.
But with today’s AI tools, a picture can look perfectly ordinary and still be generated, edited, or ripped out of context. Before you treat it as evidence, the most important step is verification. Where did the image come from? Who posted it first? And could it have been altered using AI?
One of the tools parents should understand right now is Nano Banana, Google’s highly realistic image-generation and editing system built into Gemini. Its capabilities raise serious concerns for families — especially when deepfake images are used to embarrass, sexualize, or falsely accuse kids of behavior they didn’t engage in.
Nano Banana is Google’s AI model for text-to-image generation and editing. Officially, it’s called Gemini 2.5 Flash Image. Gemini can generate extremely convincing images from a prompt and edit real photos in ways that blend smoothly, like swapping backgrounds or details, inserting or removing objects, and retouching.
Learn which AI apps your child is using with BrightCanary monitoring for iOS.
Nano Banana images look almost identical to real images. A photo can be pulled from a profile, a team site, a yearbook page, or a friend’s camera roll, and then edited to embarrass, sexualize, or frame a child as doing something they never did.
Even when an image is later proven to be an AI-generated deepfake, the social damage can stick — especially in middle school, where reputations move faster than corrections.
Parental controls for Nano Banana are the same parental controls for Gemini apps. Currently, there is no separate switch that only disables image generation.
In Family Link, Gemini access on a supervised account is essentially on or off. For kids under 13, or your country’s age threshold, a parent has to enable access before the child can use Gemini apps, and you can turn it off at any time.
Google says minors signed into their Google account have added content filters, and Gemini attempts to block categories like sexually explicit material, violence, harassment, and some harmful roleplay. But they also explicitly say these filters are not perfect.
The most practical form of protection is to replace “I saw it so it’s true” with “Don’t trust. Verify” and teach a vital life skill. Make it safe for your child to come to you without losing their phone or being grounded on the spot, because fear of punishment drives secrecy.
If there are threats, coercion, or sexualized content, preserve evidence and escalate the issue through the platform, the school, or local/federal law enforcement.
Yes, if it is an image generated by Nano Banana.
Google’s main identifier is SynthID, an invisible watermark designed to remain detectable after common changes like resizing and compression. Currently, SynthID is only used by Google, so Gemini wouldn’t be able to detect images generated elsewhere, like ChatGPT. Google does plan to implement AI image detection in search results, but there’s no target date for that yet.
So, how do you identify AI images? The SynthID page describes two ways to check. Both involve uploading the image to either Gemini or SynthID Detector portal (currently in beta) and asking if there is a watermark. The AI will tell you if it’s generated or edited by Gemini.
What if you only have a screenshot? SynthID can still work if the image is not heavily degraded or modified. Cropping the single image into smaller pieces has been reported to aid in detection.
If SynthID is detected, that is compelling evidence the content was made or edited with Google’s tools. If SynthID is not detected, that’s still not conclusive that it is real. The image could be from a different AI tool, or it could have been altered after creation or reposted enough that detection is harder. Feeding the image to another AI (i.e., taking an image from Nano Banana and input it into Grok or Midjourney) will likely not keep the watermark though regeneration.
For example, here is a SynthID check on an image generated in Gemini:
And here is the SynthID check after the image went through ChatGPT:
It’s an AI-generated photo, but the SynthID only identified the image generated with Gemini.
Remember: Visible watermarks may appear on some images, but you should never treat the absence of a visible mark as proof an image is real. Verifying can help prevent arguments, and it demonstrates that you trust your child.
With Nano Banana creations, the most common visual giveaways are not usually cartoon mistakes like weird eyes and hands. They are small failures in fine details or background elements.
Finally, ask whether the scene makes sense as a real moment. Do the setting, timing, clothing, and context line up with what you know? Ask who posted it first, whether anyone has the original post or file, and whether there are independent photos or videos from the same moment. Real events usually produce more than one angle, while a fake often lives as a single screenshot with a lot of emotion wrapped around it.
Add up what makes sense, and then decide how long they are grounded for the rest of their life.
A quick example to close with: if I saw a picture of my youngest child and saw they had matching socks? No way. That is 100% fake. Those are the kinds of details to look for.
Learn about the risks of AI and brain rot, and save these tips to help your child use AI responsibly. Monitor every app your child uses, including AI apps like Gemini and Grok, with BrightCanary.
In late December 2025 and early January 2026, multiple outlets and watchdogs reported that people were using Grok’s new image generation features to create and share nonconsensual sexualized images, including images involving children and young teens. Let’s lay out what you need to know as a parent.
Grok is an artificial intelligence chatbot built into X (formerly Twitter) and tied to Elon Musk’s xAI. It’s also a standalone app and website.
Here’s what its image features can do:
These deepfake threats can affect any family because they use the same things every teen already has: photos and social accounts.
In the past, making realistic abusive imagery took skill, time, exploiting bugs, and usually private tools. When a popular app bakes it in, the harm becomes easier and faster.
A teen can be pulled into deepfake dangers, despite having done nothing. For example, a classmate can generate or edit an image of them and share it, or a fake image can circulate in group chats, direct messages, or public posts, and pull your teen into the fallout.
This is the same emotional mechanism as sextortion and deepfake harassment — humiliation, plus panic, plus “everyone will see it.”
You do not need to become an expert on AI. You need a plan.
The safest assumption is that any photo posted publicly can be misused. The goal is not to scare them into isolation; it is to teach smart sharing. Get to the heart of what matters because shame and fear are what keeps teens quiet.
Here’s an example of what you can say to your teen:
“There’s a new wave of AI tools that can mess with photos and make fake sexual images, even of kids. If you ever see something like that, or if someone uses your photo, you’re not in trouble. Bring it to me. We’ll report it and handle it together.”
If any account your teen does not know comments, tags, or DMs about this content, tell them not to reply. This rule applies whether the photo is real or a fake image someone made. Your teen should screenshot it, block the account, report it, and tell you.
In the United States, the National Center for Missing and Exploited Children runs the CyberTipline for reporting suspected child sexual exploitation. You can also report to local law enforcement, and to the FBI (either through tips.fbi.gov or by contacting your nearest FBI field office). If there is an immediate threat or your child’s safety is at risk, call 911.
NCMEC’s Take It Down service helps remove sexually explicit images or videos depicting minors from participating platforms.
The concern at hand is not about whether Grok is “edgy,” but about whether a mainstream platform’s built-in AI image features can be used to generate and spread nonconsensual sexualized imagery, including material involving minors. You do not need to panic, but you do need a plan, and you need your teen to know they can come to you immediately if something happens.
Learn more about what to do if you find something inappropriate on your child’s phone, and stay informed about your child’s online activity with BrightCanary monitoring.
