Deepfakes, Fake Images and AI Manipulation
Deepfakes are images, videos or audio recordings created or altered using AI so that they appear real.
Some are made for entertainment, satire or creative projects. They become dangerous when used to:
- Spread lies
- Embarrass people
- Bully
- Manipulate emotions
- Create rumours
- Impersonate somebody
- Commit fraud
- Threaten or blackmail
For young people, this matters because photographs and voice recordings can be copied and reused without permission.
A normal photograph may be altered to create something the person never did. A familiar voice may be copied to make a false request sound believable.
The purpose of this lesson is not to make you distrust everything online. It is to help you recognise possible manipulation, reduce risk and know what to do if synthetic content is used against you or somebody you know.
What Is a Deepfake?
A deepfake is content created or altered using AI to make it appear that somebody said or did something that never happened.
Deepfakes may include:
- Face swaps
- Altered facial expressions
- Artificial mouth movements
- AI generated photographs
- Voice clones
- Fake telephone calls
- Edited videos
- Synthetic sexual images
Not every AI generated image is a deepfake. A picture of a fictional landscape does not pretend that a real event happened.
The greater danger arises when AI is used to imitate a real person, fabricate evidence or deliberately mislead an audience.
In the past, a photograph, video or voice recording was often treated as strong evidence. Today, digital content requires more careful checking.
How Synthetic Content Is Created
AI systems can learn patterns from photographs, video and audio.
Possible source material includes:
- Selfies
- Social media videos
- Profile photographs
- Livestreams
- Voice messages
- Gaming chats
- School or sports recordings
- Public videos
Some tools can change one face into another, copy a voice or generate a completely fictional person.
Convincing results once required specialist equipment and technical skill. Many simpler tools are now widely available.
This does not mean that posting any photograph will automatically lead to a deepfake. Most people will never be personally targeted.
It does mean that a photograph or voice recording can potentially be used in ways that were not possible before.
Deepfakes Can Spread Lies
A fabricated video can make it appear that somebody made an offensive comment, admitted to something or behaved in an embarrassing way.
People often react to dramatic content before checking whether it is genuine.
Even when a fake is later exposed, screenshots, copies and rumours may continue circulating.
Before believing a surprising recording, ask:
- Where did it first appear?
- Who uploaded it?
- Is the original available?
- Has a reliable source confirmed it?
- Could important context be missing?
- Is somebody trying to create anger or panic?
The more shocking the content, the more important it is to slow down.
Bullying and Humiliation
Deepfakes may be used to embarrass pupils or damage friendships.
Somebody might:
- Place a classmate’s face into a humiliating video
- Create a false voice recording
- Alter a photograph
- Make a fake account
- Fabricate a conversation
- Share the material through school groups
The person who created or distributed the fake may claim that it was only a joke.
It stops being a harmless joke when it frightens, humiliates or harms somebody.
Forwarding the content increases the damage, even if the person forwarding it did not create it.
If you receive harmful material, do not share it.
Synthetic Sexual Images
AI tools can create false nude or sexual images from ordinary photographs.
This is a serious form of abuse.
The victim may never have taken or sent an intimate photograph. Their face or body may have been altered or placed into artificial content.
A fake image can still cause real distress and reputational harm.
If the image appears to involve somebody under 18:
- Do not download it
- Do not take a screenshot of the image itself
- Do not forward it
- Do not joke about it
- Leave or close the content
- Report it to the platform
- Tell a trusted adult immediately
Save the account name, message, link and other nonsexual evidence where it is safe to do so. An adult can help decide what should be preserved and reported.
AI generated sexual images of children are treated as child sexual abuse material by UK child protection authorities, even when no real sexual photograph originally existed.
Voice Cloning
Voice cloning uses AI to imitate somebody’s voice.
Recordings may come from:
- Public videos
- Gaming conversations
- Voice messages
- Livestreams
- Social media posts
A false voice recording might be used to create:
- A fake confession
- An invented insult
- A false apology
- A request for money
- An urgent request for a login code
- A supposed emergency
People naturally trust familiar voices, which makes this form of impersonation effective.
If a voice message makes an unusual or urgent request, verify it through another method.
Contact the person using a number or account you already trust. Ask a question that a stranger would be unlikely to answer correctly.
Do not use contact details supplied in the suspicious message.
Fake Calls and Family Emergencies
A scammer may imitate a friend or relative and claim that something serious has happened.
The caller may demand:
- Money
- Account information
- A verification code
- Secrecy
- Immediate action
The emotional pressure is part of the trick.
If you receive a frightening call:
- Pause
- Do not send money or information
- End the call if necessary
- Contact the person directly
- Tell a trusted adult
A genuine emergency does not prevent you from checking what is happening.
Families may also agree on a private word or question that can be used to verify an unusual request.
Fake Profiles and Relationships
AI can create realistic faces belonging to people who do not exist.
A fake account may include:
- An artificial profile photograph
- AI written captions
- Fabricated interests
- Copied comments
- Stolen images
- Synthetic voice messages
This may be used to build a friendship, begin a relationship or gain trust before asking for money, information or images.
Be cautious when somebody:
- Becomes emotionally close very quickly
- Provides changing personal details
- Requests secrecy
- Pressures you for photographs
- Avoids reasonable verification
- Makes excuses whenever questions arise
- Wants to move immediately into private chat
No single photograph, video or voice recording proves that an online identity is genuine.
Fake Celebrity and Influencer Content
Deepfakes may imitate celebrities, athletes, creators or public figures.
The content may promote:
- Competitions
- Investments
- Products
- Giveaways
- Political claims
- Health advice
- Money making schemes
A familiar face does not prove that the person supports the message.
Before acting, check the person’s verified or official accounts and look for confirmation from reliable sources.
Be especially cautious when the content asks you to send money, disclose personal information or act before an offer expires.
Signs of a Possible Deepfake
Some lower quality deepfakes may contain:
- Poorly matched lip movement
- Strange lighting
- Unnatural facial movement
- Flickering hair or glasses
- Sudden changes in image quality
- A voice lacking normal emotion
- Unusual pauses or pronunciation
However, these signs are becoming less reliable as technology improves.
A convincing deepfake may contain no obvious visual mistake.
Do not rely only on trying to spot glitches. The source, purpose and surrounding evidence are often more important.
Ask:
- Is the account genuine?
- Is the request normal?
- Can the claim be confirmed elsewhere?
- Is the content trying to provoke a rapid reaction?
- Has the clip been removed from a longer recording?
- Is there an original version?
Digital investigation is more than examining a person’s face.
Emotional Manipulation
Deepfakes are effective because they can create strong emotions.
A fake may be designed to cause:
- Shock
- Fear
- Anger
- Embarrassment
- Urgency
- Sympathy
Strong emotion can make people share content or follow instructions without checking.
When something creates an immediate reaction, pause before doing anything.
Do not allow another person’s urgency to control your decision.
A few minutes spent verifying a claim may prevent a serious mistake.
Reducing Your Risk
Nobody can guarantee that their image or voice will never be copied. A victim is not responsible simply because they previously posted a normal photograph.
However, some precautions can reduce public access to personal material:
- Keep personal accounts private
- Review followers regularly
- Limit who can download or share content
- Control who can tag or mention you
- Avoid public posts revealing your routine
- Be cautious during calls with strangers
- Do not send private images
- Report copied accounts
Check the background of photographs and videos for personal information.
You do not need to disappear from the internet. The aim is to understand who can access your content.
What to Do If a Deepfake Targets You
If somebody creates false content involving you, remember:
- The fake is not your fault
- You are not responsible for the creator’s behaviour
- You do not need to face it alone
- Asking for help quickly is the right response
Try not to argue publicly with the person or repeatedly share the fake while attempting to prove that it exists. This may spread it further.
Instead:
- Stop responding to threats
- Save usernames, messages, dates and links
- Report the account and content
- Tell a parent, carer, teacher or safeguarding professional
- Ask friends not to share it
- Change account security if somebody has gained access
- Involve the police if there are threats, blackmail or sexual material
Do not pay a blackmailer. Payment does not guarantee that the content will be deleted and may lead to further demands.
Getting Sexual Images Removed
Young people in the UK who are under 18 can use Childline and the Internet Watch Foundation’s Report Remove service to report sexual images or videos of themselves and seek their removal from the internet.
Other countries and territories may offer different reporting services. A trusted adult, school safeguarding professional or police officer can help identify the appropriate route.
If the content is sexual, do not keep copying or forwarding it while seeking help. Preserve links and messages without creating additional copies of the image.
If You Receive a Threat
A person using a deepfake for blackmail may say:
“Pay me or I will share it.”
“Send another image.”
“Give me your password.”
“Do not tell anybody.”
Do not follow these instructions.
Take the following steps:
- Stop replying
- Do not pay
- Do not send more content
- Preserve safe evidence
- Block and report the account
- Tell a trusted adult immediately
Blackmail depends on fear and isolation.
Telling somebody removes the secrecy the offender is trying to create.
Supporting Somebody Else
If somebody else is targeted:
- Do not share or request the content
- Do not joke about it
- Report it
- Ask others to stop forwarding it
- Support the person privately
- Encourage them to tell a trusted adult
- Tell an adult yourself if safety is at risk
Avoid asking the victim to prove what happened by sending you the image.
A useful message might be:
“I believe you. This is not your fault. Let’s get help.”
One calm and supportive response can make somebody feel far less alone.
Protecting Mental Wellbeing
Being targeted by a deepfake may cause:
- Panic
- Shame
- Anger
- Fear
- Anxiety
- Difficulty attending school
- Loss of confidence
These feelings are understandable, but the responsibility belongs to the person who created or shared the content.
If the situation is affecting sleep, school or mental health, speak to a trusted adult or appropriate support professional.
If you feel in immediate danger or at risk of harming yourself, contact emergency support or tell an adult straight away.
No online image or rumour should be faced alone.
Final Thoughts
Deepfakes can make false images, video and audio appear real.
They may be used for entertainment, but they can also support bullying, impersonation, fraud, sexual abuse and blackmail.
Visual mistakes are no longer a reliable way to identify every fake. Checking the source, verifying unusual requests and questioning emotional content are increasingly important.
If synthetic content targets you, do not pay, negotiate or remain silent. Preserve safe evidence, report it and involve a trusted adult.
A fake image can cause genuine harm, but it does not change the truth about the person being targeted.
Technology is becoming more convincing. Calm judgement, verification and support remain powerful forms of protection.