synthetic media

basicwidespread

An image, video, audio, or text that was partially or fully generated or altered by artificial intelligence, not by a camera, microphone, or human hand.

In plain terms

Synthetic media is a collective term: it includes any image, video, audio, music, or text created or modified by artificial intelligence. This includes photorealistic images made from a few-word description, audiobooks read by machines, advertising characters with non-existent faces, or videos in which a real person says something they never said.

Think of it as synthetic fiber in the textile industry. Cotton grows in the fields, polyester is made in a factory, yet both end up as clothes. Synthetic media is also "manufactured" content: not an imprint of reality, like a photograph, but the result of a model's calculations. It can still be useful, beautiful or misleading.

The term itself is neutral. Synthetic media includes film dubbing where the actor's mouth aligns with the new language, artificially restored voices for patients who lost their speech, and cheaply produced training videos. The same technology underlies scammers' voice cloning and political deepfake videos.

It differs from traditional trick photography and photo retouching in that it requires no studio, expertise, or much time. With a few sentences of instruction, anyone can create convincing content in minutes.

How it works

Training: a generative model processes vast amounts of real images, audio recordings, videos, or text and learns their patterns, such as how light falls on a face or how a human voice fluctuates.
Instruction: you specify what you want. This can be a text description (prompt), a sample image, or a few seconds of audio if someone's voice needs to be imitated.
Generation: the model computes new content from learned patterns; it does not splice together existing recordings. Images and video are typically produced by gradually denoising random noise, while text is written piece by piece, one after another.
Refinement: the finished material can be further modified, for example by face swapping, aligning mouth movements with the text, rewriting the background, or synchronizing audio and video. Increasingly, an AI agent stitches these steps together with little or no human intervention.
Labeling: responsible providers embed an invisible watermark or provenance data in the file so that the machine origin can be demonstrated later. However, this is not available in every tool and can be removed.

An everyday example

On your social media, a well-known Hungarian doctor 'recommends' an arthritis cream advertised as a miracle cure in a video. The voice and face are the doctor's, but the doctor never said this: the video was produced with artificial intelligence from an earlier interview. It is synthetic media, but it can also be harmless, like the machine voice that reads out navigation instructions and street names in your car.

Who is affected

everyoneYou can no longer tell at a glance whether what you see or hear online actually happened, so checking the source has become more important than ever.

marketerYou can create ad images, dubbing, and product videos at a fraction of the previous cost, but in many cases you must label the machine origin, and you cannot use the likeness of a real person without permission. → Marketing

bank customer, investorScammers can impersonate your relative, boss, or bank officer using cloned voices or fake video calls. → Money

parent, student, teacherChildren easily create and receive fake images of each other, so conscious media use must be discussed both at home and in school. → Education & Family

A common misconception

MythSynthetic media is the same as deepfake, meaning it is always a deception.

In realityDeepfakes are only the deceptive slice of synthetic media. Most machine-generated content is entirely ordinary and lawful: read-aloud text, illustrations, dubbing, educational videos.

Limits and risks

It is increasingly difficult to detect with the naked eye and ear, and detection programs often make mistakes in both directions.
Watermarking and provenance labeling are useful but not proof: their absence does not mean the content is real, because the label can be removed or may be missing from the start.
Unauthorized use of a real person's face or voice usually violates personality rights and can constitute a crime in cases of fraud or harassment.
As a side effect, the credibility of real footage weakens: anyone can claim that an awkward but true video was 'just made by the machine'.

Similar, but not the same

deepfake – A deepfake is machine-generated content that deceptively imitates a real person or event, while synthetic media covers all machine-generated content, including harmless ones.
synthetic data – Synthetic data is an artificially generated dataset for training or testing models, not content intended for humans to view or listen to.
computer animation (CGI) – Traditional CGI is built by experts step by step, while synthetic media is generated by a trained model from a short prompt.

Related concepts

Broader

generative AI, digital media

Narrower

deepfake, voice cloning, image generation, video generation, text-to-speech

On the edge

watermarking, origin authentication, diffusion model, disinformation, digital avatar

A brief history

One technical foundation of today's synthetic media is generative adversarial networks (GANs), introduced in 2014, while the term 'deepfake' spread in 2017. Tools that generate images from text prompts, released in 2022, made the technology accessible to the masses.

Where things stand

By the end of 2026, AI-generated images, audio and video have become an everyday part of advertising, entertainment and social media, often at a quality indistinguishable from the real thing. The transparency and labelling rules of the EU AI Act (Article 50) apply from August 2026, but labelling in practice is still not uniform.

Status: 2026 November

In the news

Appears in 35 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Meta makes advertisers solely responsible for its AI changes to ads; new API features arriveNov 2, 2026 · Marketing
Etsy restricts AI-generated content to curb seller slopOct 23, 2026 · Marketing
Hungarian universities' AI rules: ELTE criterion course, MATE plagiarism rule, PTE interviewOct 19, 2026 · Education & Family
Axios: AI slop websites up fivefold as GEO optimization spreadsOct 17, 2026 · Marketing
AI ad labelling: responsibility stays with the advertiser, while the EU requires transparencyOct 12, 2026 · Marketing
Stack Overflow relaunches Developer Story to filter AI-noisy recruitmentOct 12, 2026 · Code
AI Act transparency rules: labelling AI content and deepfakesOct 12, 2026 · Jog
Podcasts address employment-law and compliance questions on workplace AI useOct 6, 2026 · Jog

Also known as

synthetic media · synthetic media · szintetikus tartalom · AI-generated content · AI-generated content · AI-generated content · AI-generated media · generative media

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.