unhackchildhood.com/topics/ai-deepfakes · Reviewed September 24, 2026
DO THIS FIRST
Take back some control.
- 01
Know what it is: generative AI predicts plausible words, images, audio, or video from patterns in training data. It does not understand or verify facts like a person does.
- 02
Pause before sharing a shocking image, recording, or claim. Find the original source and confirm it through reliable reporting.
- 03
Verify urgent voice or video requests through a separate, known contact method.
- 04
Keep names, school details, passwords, health information, private photos, and other sensitive material out of prompts.
- 05
Save evidence and report fake sexual, humiliating, threatening, or impersonating content.
- 06
Check the school or teacher’s rules before using AI for an assignment, and say how it helped when disclosure is required.
FOR STUDENTS
No shame. Just the next move.
- Good uses include brainstorming questions, practicing a skill, explaining a hard idea in another way, and giving feedback on work you already did.
- Bad uses include outsourcing the thinking, inventing sources, impersonating someone, making sexual or humiliating images, and treating a chatbot like a confidential counselor.
- Ask for sources, then open and check them. AI can confidently invent facts, quotations, citations, and links.
- You are responsible for what you submit, create, or share—even when a machine made the first draft.
FOR PARENTS
Contain first. Debrief later.
- Try the tools together before allowing independent use. Check the provider’s current age rules, privacy terms, history controls, and parental settings.
- Make three family rules explicit: no sensitive data in prompts, verify important answers elsewhere, and bring upsetting or sexual output to an adult.
- Keep companion-style chatbots out of the role of therapist, romantic partner, or secret friend. Their human-like tone can create trust they have not earned.
- Build a family verification habit: pause, call back on a known number, and use a private family phrase for emergencies.
- Discuss consent before children use someone’s likeness or voice in a generated joke.
- Treat synthetic bullying and sexual content as serious even when someone says, “But it isn’t real.” The harm is.
TRUSTED NEXT STOPS
Official help and reporting.
Reviewed September 24, 2026. External guidance changes; confirm time-sensitive steps on the linked official site.
SET IT UP