What the term covers
Synthetic media is a broad label for images, video, audio or text that has been produced or heavily altered by software rather than captured directly from the real world. “Deepfake” usually refers to a narrower and more worrying case: a realistic fake that shows a real person saying or doing something they did not say or do, or that swaps one person’s face or voice onto another.
Not all synthetic media is a problem. Animation, visual effects, dubbing and illustration have long used artificial techniques, and generated material can be used openly and creatively. The concern is deception, and especially deception about real people.
How fabricated media can mislead
A fake does not have to be perfect to cause trouble. Several effects tend to appear:
- False events. A clip of a public figure appearing to say something they never said can spread before anyone checks it.
- False identities. A cloned voice or face can be used to impersonate someone, including in scams aimed at friends or relatives.
- The liar’s dividend. When fakes are common, real evidence can be dismissed as “probably fake.” That damages honest reporting as well.
- Harm to individuals. Fabricated intimate, humiliating or defamatory material can seriously hurt the person depicted, whether or not viewers believe it.
- General erosion of trust. Even debunked fakes can leave people less sure about what they see and hear.
Audiences often form an impression within seconds, and corrections tend to travel less far than the original. That is one reason prevention matters more than cleanup.
What careful creators avoid
You do not need to be a technician to follow a few clear boundaries.
Do not depict real people doing or saying things they did not. This applies even when you consider it a joke. What seems obviously silly to you may be taken seriously by someone who sees a clipped version. Personal likeness deserves particular care.
Do not clone a voice or face without permission. Consent to feature someone in a recording is not the same as consent to simulate them. If a person has not clearly agreed to a synthetic version, leave it out.
Do not present generated scenes as documentation. An invented image of a news event, a historical moment or a disaster, presented as real, is misinformation, even if you did not intend harm.
Do not use realistic fakes to pressure or deceive. This includes fake endorsements, fake testimonials and fake “before and after” material.
Do not create sexual or degrading depictions of real people. Such material is harmful and, in many places, unlawful. Check official guidance for your location.
Labeling and context
When synthetic elements are a legitimate part of your work, say so clearly. A label should be easy to see at the point of viewing, not buried in a long description. The reasoning behind that is covered further under AI disclosure.
Useful context includes:
- A plain statement that the material is generated or altered
- Clear framing as satire, fiction or illustration where that applies
- Notes on which parts are real and which are not
Labels help, but they are not a license for harm. A fake that damages a real person remains a problem even if a disclaimer sits in small text beneath it.
Reasons a creator might use synthetic media responsibly
There are legitimate uses where nobody is deceived, for example:
- A fictional character voiced by a generated voice that does not imitate a real person
- A reconstruction clearly marked as an illustration of what an event might have looked like
- A performer who has agreed, in writing and with clear terms, to a synthetic version of their own voice or appearance
- Visual effects in a story that audiences understand to be fictional
In each case the questions are the same: Who is represented? Have they agreed? Will viewers understand what they are seeing?
Practical checks before sharing
You may also encounter other people’s media and be tempted to repost it. A brief pause helps.
- Look for the original source. Who first published it, and where? Clips with no traceable origin deserve more doubt.
- Check whether reputable sources are reporting it. A startling claim reported nowhere else is a warning sign. This is part of the habit of self fact-checking.
- Notice small oddities. Unnatural blinking, mismatched lighting, garbled text, odd hands or audio that does not quite fit the mouth can be clues, although absence of such flaws proves nothing, and the tools keep improving.
- Consider the emotional pull. Content designed to provoke outrage or fear quickly is more likely to be shared unchecked.
- When uncertain, do not share it as fact. Saying “I could not verify this” is more honest than repeating a claim.
No single check is reliable. Treat them as ways to slow down, not as tests that give certain answers.
If you are the target
If someone creates a fake of you, keep copies and note where it appears, then use the reporting tools of the platforms involved. Official guidance from consumer protection or privacy bodies in your area can explain options, and for serious cases a qualified legal adviser may help. Outcomes vary, and removal is not always quick.
Why this matters for creators
Public trust in what is shown is a shared resource. Every creator who labels material honestly and refuses to fabricate real people helps keep that resource intact. Working with generated media responsibly generally means choosing clarity over cleverness, and treating the people who appear in your work, even in simulated form, as people rather than assets.