Settings & Vocab

Safety settings & AI words, explained

Ten minutes here will make every other conversation easier.

Settings to check in every AI app

From Brookings[3]' tip sheet on kids' data:

  • Turn off using your child's data to train AI models (often called “help improve our model”).
  • Turn off using past chats and “memories” to personalize ads.
  • Turn off location collection.
  • Read the privacy policy of anything that records, listens to or tracks your child, including smart speakers, watches and location trackers.
  • Look for alternatives that collect less data.

Parental controls: what they do (and don't)

Teen accounts & parent linking

Some tools, like ChatGPT and Gemini, let parents link accounts and adjust safety settings (Internet Matters[8]).

What they don't do

Controls are not a force field. Independent testing of ChatGPT's controls found gaps (FOSI[9]), and age checks on many apps are easy to get around (Internet Matters[8]). Conversations still matter most.

What good AI should have

The APA[4] calls for age-appropriate defaults, reminders that you're talking to a bot, less persuasive design, and easy ways to get human help.

AI words in plain English

Switch to the kid version for simple explanations to share with your child. Terms meant for parents only are hidden in the kid version.

Generative AI
Technology that creates new text, images, audio, video or code after being trained on huge amounts of data from books, websites and more. It doesn't actually “know” or “think.”[1]
A computer program that makes new words and pictures by copying patterns it has seen.
Chatbot
A program that writes back to you like a person, often designed to feel human and emotionally responsive.[2]
A computer program you can type to. It sounds like a person, but it isn't one.
AI companion
A chatbot designed to act as a friend or romantic partner. These encourage frequent, personal chats and have some of the weakest privacy protections.[3]
A chatbot that pretends to be your friend. Real friends are people.
Hallucination
When AI states something false with confidence.
When AI makes something up and sounds sure about it.
Deepfake
A fake photo, video or voice made with AI that looks or sounds real.[4]
A fake picture or video that a computer made to look real.
Image-manipulation app
A tool that uses AI to alter photos of real people, sometimes to create harmful or inappropriate fake images. Creating or sharing such images of minors is illegal in many places.[5]
Image-based blackmail
Threats to share a private or AI-altered image of someone unless they pay or comply. Payment usually leads to more demands.[6]
Voice cloning
Copying someone's voice from a few seconds of audio, used in fake “emergency” scam calls.[7]
When a computer copies a voice. That's why we have a family safe word.
Data harvesting
When AI tools automatically gather information like location, health details and what kids say, building a detailed profile over time.[3]
When apps collect little bits of info about you, like puzzle pieces.
Training data
The material an AI learned from. Some apps also train on users' chats unless you turn that off.[3]
All the stuff the computer read so it could learn.
Prompt
What you type or say to ask AI to do something.
The question or instructions you give AI.
Persuasive design
Features built to keep you using an app, like constant notifications, rewards or personalized replies.[4]
Tricks apps use to keep you playing longer.
COPPA
The U.S. law requiring parental consent before collecting data from kids under 13. No similar federal consent rule covers ages 13 to 17.[3]
A rule that protects kids under 13 online.
Age gate
A check meant to keep younger kids out of an app, often just typing a birth date.[8]
The “how old are you?” question. Always tell the truth.

References

AI-OK? summarizes these sources in our own words and links to the originals. Short quotes are marked with quotation marks, and some titles are described in [brackets]. All credit for original research and guidance belongs to the authors below.

  1. Brookings Institution. (2026). What is generative AI and what's at stake for kids? Parent tip sheet. https://www.brookings.edu/wp-content/uploads/2026/01/Brookings_What-is-AI_Flyer_DIGITAL.pdf. Accessed October 4, 2026.
  2. Wolbert, E. D., Rudy, T. L., & Perkins, D. F. (2026, March 18). AI chatbots and children's digital safety. Penn State Extension. https://extension.psu.edu/ai-chatbots-and-childrens-digital-safety. Accessed October 4, 2026.
  3. Brookings Institution. (2026). Where kids' information goes in an AI world. Parent tip sheet 11. https://www.brookings.edu/articles/tips-for-parents-raising-resilient-learners-in-an-ai-world. Accessed October 4, 2026.
  4. American Psychological Association. (2025, June). Artificial intelligence and adolescent well-being: An APA health advisory. https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-ai-adolescent-well-being.pdf. Accessed October 4, 2026.
  5. American Academy of Pediatrics. [Q&A on laws and policies around AI-generated deepfakes]. www.aap.org (link). Accessed October 4, 2026.
  6. Thorn. [Guidance for young people facing online image threats]. www.thorn.org (link). Accessed October 4, 2026.
  7. KPRC Click2Houston. (2026, June 2). FBI warns of AI voice-cloning scam that mimics loved ones in distress. https://www.click2houston.com/news/local/2026/06/02/fbi-warns-of-ai-voice-cloning-scam-that-mimics-loved-ones-in-distress/. Accessed October 4, 2026.
  8. Internet Matters. A guide to artificial intelligence (AI) for parents. https://www.internetmatters.org/resources/parent-guide-to-artificial-intelligence-ai-tools/. Accessed October 4, 2026.
  9. Family Online Safety Institute. Testing OpenAI parental controls: An evaluation of safety, gaps, and policy implications. https://fosi.org/policy/testing-openai-parental-controls-parental-controls-an-evaluation-of-safety-gaps-and-policy-implications/. Accessed October 4, 2026.