Risks
Deepfake
A video, image, or audio clip that AI has altered or generated so it looks or sounds as if a real person said or did something they did not.
- Parents
- Educators and school leaders
- Policymakers and staff
What parents should know
A deepfake is a fake recording of a real person. It can be a face in a video, a voice on a phone, or a picture that never happened. Some uses are jokes. Some are fraud or humiliation, including fake intimate images of students. Do not forward a harmful one. Check back with the real person on a number you already trust.
On this page
What is a deepfake?
Deepfake is the everyday word for synthetic media aimed at a real identity. The clip is built so viewers trust their eyes and ears. Face swaps, voice clones, and fully generated scenes are the common forms. A filter that adds cartoon ears is not what people mean. A clip of a principal saying words they never said is.
The file does not have to be perfect to do harm. A rushed fake in a group chat can spread before anyone checks. Content credentials and watermarks try to label the file. Most clips a family will see this year will not carry a mark you can rely on. The check is still a second channel: call the person, ask the school, look for the same scene in a real news source.
Why a deepfake matters
Students are both the audience and the target. A fake fight video, a fake confession, or a fake nude can move through a grade in an afternoon. Schools need a rule that says do not forward, support the person pictured, and do not display the file to prove the point.
Parents meet a second version on the phone: a child's voice, crying, asking for money. Hang up and call your child on the number you saved. An AI impersonation scam depends on you not taking that extra minute.
How it shows up in practice
- A group chat shares a clip of a student that the student says is fake, and a parent does not resend it.
- A voicemail sounds like a child in trouble, and the parent calls the child's real phone before sending money.
- A school tells staff not to project a suspected deepfake in a meeting.
- An intimate fake of a minor is reported to NCMEC instead of being saved to more devices.
How HeyOtto helps
HeyOtto does not offer face swap, voice cloning of a real person, or tools that turn an ordinary photo into a nude. Image generation is age-filtered. A public share of a child's project needs parent approval and can be revoked. Parents can read the chat that asked for an image. None of that removes a deepfake made somewhere else. If a minor is in an intimate fake, report it at report.cybertip.org and use NCMEC's Take It Down. Do not forward the file.
- Schools can turn image generation off.
- Otto is the assistant. It is not a tool for impersonating a classmate.
- HeyOtto does not claim a watermark on images made outside the product.
For families
Try freeFAQs
How can I tell a video is fake?
You often cannot, from the clip alone. Call the person on a number you already have. Ask whether a news outlet you trust is reporting the same scene. A watermark may be missing even when the clip is fake. Do not decide by how emotional the video feels.
What if the deepfake is an intimate image of a child?
Do not forward, post, or inspect it out of curiosity. Report to NCMEC's CyberTipline. NCMEC's Take It Down service helps with removal of intimate images of people under 18. The TAKE IT DOWN Act also requires covered platforms to remove a validly reported intimate image within 48 hours.
Should my child send the clip to friends to warn them?
No. Warning people by forwarding the file is how it spreads. Tell them it is fake, without the attachment. If classmates already have it, tell the school and ask them not to display it.
Are all AI images deepfakes?
No. Synthetic media includes harmless filters and drawings of imaginary scenes. A deepfake points at a real person and a real-looking event or statement that was not real. The risk depends on whose identity is in the file.
Sources
Last reviewed September 26, 2026. This entry is reviewed twice a year.