This article explains what deepfakes are, how they harm victims, what to do immediately, and how families can reduce risk.
Why This Matters Now
Deepfakes are no longer a niche internet trick. They are being used for fraud, impersonation, sexual abuse, extortion, and misinformation, and they are getting cheaper and easier to produce. Officials in Australia, Singapore, the United States, the United Kingdom, the European Union, and UNICEF are all treating the problem as a real public-safety issue, not just a tech curiosity.
This matters most for ordinary people. You do not need to be famous to be targeted. A few public photos, short video clips, or a voice sample can be enough for criminals to build fake media that looks or sounds convincing enough to pressure relatives, embarrass victims, or steal money.
What Deepfakes Actually Are
A deepfake is synthetic media created or manipulated with AI to make it appear that a real person said, did, or appeared in something that never happened. That can mean a cloned voice, a face-swapped video, an AI-made sexual image, or a fabricated “proof” clip used in a scam.
The key point is not the model type. The key point is the effect: deepfakes are built to trick people into believing false content is real. That is why the threat is not limited to politics or celebrities. It reaches families, schools, workplaces, and private relationships.
“Deepfake abuse is abuse.”
That line from UNICEF matters because it cuts through the nonsense. There is nothing harmless about fake intimate images, fake emergency calls, or fake videos used to humiliate a child or blackmail an adult. The media may be synthetic, but the damage is real.
How Deepfakes Harm Real People
Deepfakes usually hit in four ways:
- Fraud and impersonation: scammers clone a voice or fabricate video to demand money fast.
- Sexual abuse and extortion: fake intimate images are used to shame, threaten, or control victims.
- Reputation attacks: false clips are pushed to damage trust, relationships, or careers.
- Child exploitation: sexualised AI-generated images of children are treated by UNICEF and child-safety bodies as abuse, not a joke or prank.
The child-safety point is critical. UNICEF has warned that AI-generated sexualised images of children mark a serious escalation in risk, and it is explicit that this material is child sexual abuse material.
Do Not Rely on Your Eyes Alone
A lot of bad advice online still says you can catch deepfakes by spotting weird blinking, bad shadows, or warped teeth. Sometimes that works. Often it does not. Singapore’s Cyber Security Agency says detection tools are still at an early stage, and CSIRO-backed research found major weaknesses in widely used deepfake detectors.
So the smarter rule is this: do not ask only “Does this look fake?” Ask “Where did this come from, who wants me to act, and how do I verify it?” That mindset is more useful than pretending ordinary people can reliably inspect every AI fake by eye.
How to Check a Suspicious Video, Image, or Voice Call
Use this order:
1. Check the source
Was it sent from the real account, number, or platform you already know? Or did it arrive from a new account, random number, or forwarded chain? Scam pressure usually starts with a bad source.
2. Check the context
Is the request urgent, emotional, secretive, or demanding money right now? That is a scam pattern, especially in cloned-voice and family-emergency frauds.
3. Verify on a separate channel
Call the person back on the number you already have. Message them through an existing chat. Contact another relative. Do not stay trapped inside the same suspicious call or message thread.
4. Look for provenance or labels
Where supported, check for visible labels, source information, or Content Credentials showing how the media was made or edited. These systems are not universal yet, but they are part of the shift from unreliable guesswork to verifiable provenance.
What To Do Immediately If You Become a Victim
This is the section that matters most.
Stop contact and do not pay
If someone is threatening you, blackmailing you, or demanding money to keep fake content offline, stop engaging and do not pay. eSafety is explicit on this point: paying usually does not solve the problem.
Preserve evidence
Save screenshots, URLs, usernames, dates, messages, and copies of the threat. Evidence matters for takedowns, platform reports, police reports, and legal escalation. Do not rely on memory.
Report the content where it appears
Use the reporting tools on the platform or service first. Fast reporting can reduce spread, create a record, and help with removal.
Use specialist removal tools
For intimate-image abuse involving adults, StopNCII can create a hash of the image so participating platforms can help detect and block sharing. For minors, NCMEC’s Take It Down is built for nude, partially nude, or sexually explicit images and videos involving people under 18. Google also has removal pathways for personal adult sexual content and artificial imagery in Search.
Secure your accounts
Change passwords, enable multi-factor authentication, review recovery options, and lock down privacy settings. If the abuser is impersonating you, warn close contacts so they do not trust sudden requests, emergency stories, or payment demands.
Escalate where needed
If the material is sexual, involves a child, includes extortion, or creates a real-world safety risk, report it to the relevant authority or child-safety body in your country. In Australia, eSafety has direct pathways for image-based abuse.
What Not To Do
A lot of victims make the situation worse by panicking. Do not do these:
- Do not pay extortion demands.
- Do not keep negotiating with the blackmailer.
- Do not repost the fake content to “expose” it.
- Do not share or download sexualised images of a minor to show other people.
- Do not assume a detector app gives you a final answer.
Prevention: Reduce the Raw Material and Reduce the Scam’s Chances
Prevention has two parts.
Reduce the raw material
You cannot disappear from the internet, but you can stop making the job easy.
- Lock down public social profiles.
- Limit public face videos and clear voice samples.
- Be selective about what you post of your children.
- Review who can tag, download, or repost your content.
- Remove old public material where possible.
Reduce the scam’s chances of success
Even if a deepfake exists, it does not have to work.
- Create a family rule: no money sent on urgency alone.
- Call back on known numbers.
- Use a family verification phrase for emergencies.
- Treat urgent secrecy as a warning sign.
- Teach children to report fast, not hide in shame.
A Simple Family Safety Checklist
| Action | Why it matters | What to do |
|---|---|---|
| Family callback rule | Stops panic-driven scams | Never act on an emergency voice or video message without calling back on a known number |
| Shared verification phrase | Helps defeat impersonation | Use a private family phrase for urgent situations |
| Privacy audit | Reduces training material | Review every public account, especially photos, videos, and children’s content |
| No-shame reporting rule | Speeds up help | Tell kids and teens they will not be punished for reporting fake nudes, threats, or strange requests |
| Account hardening | Limits hijack and impersonation damage | Use strong unique passwords and MFA |
| Safer search settings for kids | Cuts accidental exposure | Use search and platform safety controls where available |
The no-shame rule matters more than most parents think. A child who fears punishment is far more likely to stay silent when a fake sexual image or blackmail threat appears. Trusted-adult reporting is one of the clearest messages from child-safety guidance.
The Law Is Catching Up, But Not Fast Enough
Governments are moving, but the patchwork is still uneven. Australia already provides reporting and removal pathways for image-based abuse through eSafety. The UK has pushed platforms toward stronger online-safety duties and is consulting on additional measures including hash-matching for intimate image abuse. The EU AI Act’s transparency rules are set to come into effect in August 2026, including requirements aimed at making AI-generated content identifiable and clearly labelling certain deepfakes.
That helps, but law alone will not protect your family day to day. Fast verification habits, fast reporting, and fast removal requests still matter more in the first hours after a deepfake appears.
Conclusion: Deepfakes Are Here. Panic Is Optional.
Deepfakes are getting better. That part is real. But the answer is not paranoia and it is not blind trust in “deepfake detector” apps. The answer is a harder, smarter routine: verify through another channel, preserve evidence, report fast, use specialist takedown tools, lock down accounts, and make sure every person in your home knows that fake abuse should be reported immediately, without shame.
The blunt truth is simple: you may not be able to stop every fake from being made, but you can make it much harder for a fake to fool your family, spread unchecked, or control your next move.