Screenwise parents frequently ask how to handle the sudden explosion of conversational AI in their children's lives. The reality is that securing platforms like ChatGPT and Character.ai requires more than just a standard screen time limit; it requires managing specific account linkages, age assurance protocols, and feature toggles. To protect your children, you need to link your teen's ChatGPT account to restrict voice and image features, while understanding how Character.ai's recent November 2025 security overhaul has restricted open-ended chat for minors. This guide walks through the exact settings and physical steps required to lock down these tools on family devices.
Navigating the Character.ai under-18 overhaul with a digital parenting lens
In late 2025, the interactive entertainment platform Character.ai faced intense scrutiny from regulators, child safety advocates, and parents regarding how young users interact with custom-built artificial intelligence personalities. The platform, which allows users to roleplay with fictional characters or historical figures, historically operated with very few guardrails separating adult and teen experiences. That changed on November 24, 2025, when the company began rolling out a sweeping policy shift designed to protect minors.
According to the official Important Changes for Teens on Character.ai – C.AI Help Center, the platform has completely eliminated open-ended chat for users under 18. This transition began with a strict daily time limit of two hours for minor accounts, which was systematically reduced to prevent long-term emotional dependency on artificial personas. If your teenager has been using this app for roleplay or companionship, they will now experience a highly restricted environment that prevents open-ended, unpredictable text exchanges.
To enforce these safety boundaries, the platform implemented a dual verification approach. It combines an in-house age-prediction model with a third-party identity verification service called Persona. If the system flags an account as belonging to a minor, or if the user self-reports an age under 18 during signup, the restricted teen model is applied automatically. This younger user experience employs strict content classifiers to block sensitive, suggestive, or romantic roleplay.
For parents, this means the days of attempting to manually monitor hours of chat logs are over because the platform has essentially removed the capability for open-ended conversational exploration for minors. However, teenagers frequently attempt to bypass these age-gate systems by creating new accounts with falsified birthdates. If you suspect your child is accessing the adult version of the platform, look for the Persona age verification prompt or check the account profile settings to verify the registered birth year.
Linking accounts in the ChatGPT parent portal: What Screenwise parents need to know
While Character.ai chose to restrict access, OpenAI took a different route by introducing built-in parental controls. These controls allow parents to link their own ChatGPT account to their teenager's account, creating a supervised framework. Unlike traditional social media monitoring tools, this system is designed to respect teen privacy while still giving parents control over specific operational features.
Before you begin the linking process, understand the boundaries of this system. As detailed in the Parental controls to shape ChatGPT for your family resource, parents do not have access to read their teen's private conversation transcripts. The system is built this way to encourage teenagers to use the tool for homework and research without feeling constantly watched.
Instead of total surveillance, the platform relies on automated safety triggers. If the AI detects acute safety risks, such as self-harm indicators or violence planning, the system alerts trained human reviewers. If those reviewers confirm a genuine risk, the parent is notified immediately via email, text message, or push notification. This approach represents a shift from the typical surveillance style of traditional platforms, which we discuss in our realistic guide to Reddit safety settings for teens.
To link your teen's account to your own, follow these steps:
- Open the ChatGPT website or mobile app and log into your adult account.
- Select your profile icon in the top-right corner, then click Settings.
- Navigate to the Parental Controls tab in the sidebar menu.
- Click "+ Add family member" to generate an invitation link.
- Send this invitation to your teenager via email or text message.
Once your teenager accepts the invitation, their account status will update to "Linked" inside your parental controls dashboard. It is important to know that teens can choose to unlink their account at any time. If they do, the system will immediately send you a notification so you can address the change directly with them.
Locking down ChatGPT's most unpredictable features on family devices
Once the accounts are successfully linked, you can customize the individual feature toggles to create a safer environment. The default teen settings already block standard adult topics, but parents can manually restrict the most interactive and unpredictable components of the platform.
Disabling voice and image generation
The mobile version of ChatGPT includes a highly realistic voice mode that allows users to have spoken conversations with the AI. While helpful for language practice, voice interactions can make the AI feel deceptively human, which sometimes leads to unhealthy emotional attachment. Inside the linked parental controls menu, you can toggle the "Voice mode" switch to "Off." This completely removes the audio chat icon from your teen's mobile application.
Similarly, you should consider disabling the image generation feature, which is powered by DALL-E. Although OpenAI uses automated filters to block the creation of explicit or copyrighted images, kids often find creative ways to bypass these prompts. Turning off the "Image generation" toggle prevents the teenager from creating or editing any visual media, limiting the tool strictly to text-based research and writing assistance.
Turning off model training and memory
By default, ChatGPT uses user conversations to train future iterations of its large language models. It also uses a feature called Memory to remember personal details across different chat sessions, such as a user's hobbies, family members, or school schedule. For teenagers, this means the AI can build a detailed personal profile over months of continuous use.
To protect your child's data privacy, you should navigate to your teen's settings under your family dashboard and turn off two specific options:
- Improve the model for everyone: This stops OpenAI from saving or reviewing your child's chat transcripts for training purposes.
- Reference saved memories: This disables the persistent memory storage, ensuring that every new chat session starts with a clean slate and the AI does not retain personal information.
Setting quiet hours and managing safety alerts through digital parenting tools
One of the most practical tools in the OpenAI parental dashboard is the ability to schedule access limits. Because AI chatbots are always available, teenagers often turn to them late at night when they are feeling lonely or struggling with insomnia. This can lead to late-night rabbit holes that disrupt sleep.
The "Quiet hours" feature allows you to establish specific blackout periods during which the tool cannot be accessed. For example, you can block usage between 10:00 PM and 6:00 AM on school nights, or during designated homework hours. If your teen attempts to open the app or send a prompt during these blocked windows, they will see a simple notification stating that ChatGPT is currently unavailable.
When configuring these schedules, talk to your teen about why these boundaries exist. Explain that these tools are meant to be assistants, not late-night companions. Make sure your notification preferences are set to send alerts directly to your mobile phone via SMS or push notifications so you are immediately aware if any acute safety reviews are triggered by the system.
Why platform controls fail and how to secure family devices
While built-in parental settings are a helpful first line of defense, they are not foolproof. The biggest flaw with any account-level control is that it requires the user to be logged in. If your teen simply opens a private browser tab and visits the website without logging in, they can chat with the basic model without any of your custom settings, schedules, or feature blocks in place.
Furthermore, kids can easily access alternative platforms like Claude, which currently offers no family accounts or parental controls beyond a simple age-verification checkbox. To prevent these workarounds, you must combine platform-specific settings with device-level safeguards.
First, consider blocking access to unauthorized AI websites at the router level or through your device's built-in operating system controls. If your child uses an iOS device, you can use Screen Time to limit adult websites, which helps block anonymous browsing access to unlinked AI tools.
Second, you need to address the browser itself. To prevent your child from bypassing these boundaries using anonymous windows, read our guide on how to permanently disable incognito mode across your child's devices. Disabling incognito mode ensures that all web traffic passes through your filtered browser settings, making it much harder for teens to use unauthorized versions of these conversational tools.
Ultimately, protecting your family in the age of generative AI requires a mix of technological boundaries and open communication. Securing the settings on ChatGPT and understanding the limits of Character.ai are excellent steps, but they work best when paired with a clear family understanding of how these tools should be used.
Take the free, anonymous Screenwise 5-minute survey to get an instant, personalized list of developmentally positive shows, games, and apps tailored specifically to your family's values by visiting Screenwise.