A video of a politician saying what they never said. A voice that deceived an elderly mother. A familiar face in a stranger's context. Deepfakes are already here — and they change not only media, but the mind.
In 2023 a woman in Spain transferred 4,500 euros to fraudsters. She was convinced: her son was calling. The voice was his. The intonations — his. The request — urgent, he was in trouble, money needed right now. The voice had been synthesized from several minutes of social media recordings.
This is not an exceptional case. This is the new normal. And it is changing a fundamental psychological mechanism that holds human society together: trust in one's own senses.
A deepfake (from deep learning + fake) is content created or altered using neural networks so that it is perceived as genuine. A video in which one person's face is replaced by another's. Audio synthesized from a few minutes of voice recording. Images of non-existent people indistinguishable from real photographs.
The technology is rapidly becoming cheaper. If in 2017–2018 a deepfake required powerful equipment and thousands of hours of footage, by 2024 smartphone apps create convincing voice synthesis from 10–20 seconds of audio, and a video deepfake in a few minutes on consumer hardware.
The human brain was not built for skepticism toward sensory input. Evolutionarily we trust what we see and hear — because in the pre-digital era the perceived always had a real source. A tree is a tree. Mom's voice is Mom.
Deepfakes exploit this evolutionary trustfulness. The brain processes audiovisual information largely automatically, at a pre-conscious level. "This face is a real person" is decided before critical thinking activates. And critical thinking requires effort — which we do not always apply, especially when emotionally engaged.
This is precisely why deepfakes are so effective in situations of emotional pressure. Fraudsters do not call on a calm evening: they call "urgently," create panic, give no time to think. Under anxiety, critical analysis is the first thing to switch off.
Researchers describe an effect they call the "liar's dividend": the possibility of deepfakes is used not only to create fakes, but also to deny the genuine. A politician caught on a real video recording claims: "This is a deepfake." Proving otherwise becomes increasingly difficult.
More broadly, the spread of deepfakes creates what sociologists call an "epistemic crisis": a situation in which society loses shared reference points for determining truth. When you cannot trust your own eyes — what remains?
Sociological research in the US and Europe records a steady trend: trust in media content is declining. But not in a differentiated way — people do not become smarter at distinguishing genuine from fake, they simply trust everything less. This is not skepticism. It is trust nihilism.
Beyond the broad social consequences, deepfakes create personal traumas. Several groups are most vulnerable.
Victims of pornographic deepfakes. According to Sensity AI, more than 90% of deepfakes online are pornographic. Almost all victims are women. Psychological consequences: severe shame, anxiety, a sense of losing control over one's own body and image, PTSD-like symptoms.
Elderly people are the primary risk group for voice fraud. It is harder for them to apply critical thinking to technology they do not know.
Public figures — politicians, journalists, activists — face targeted discrediting attacks through fake videos.
Full "deepfake literacy" is a skill that develops over time. Several anchors:
Emotional pause. If content triggered a sharp reaction — shock, anger, fear — this is a signal to stop. It is exactly in states of emotional arousal that we are most vulnerable to manipulation.
Source verification. Where does this video/audio come from? Who published it first? Are there other sources of the same event?
Technical markers. Deepfakes often have artifacts: unnatural blinking, distortions at the boundary of face and hair, lip-audio desync, strange lighting. The technology is improving, but so are detectors.
For businesses and organizations: introduce identity verification protocols through additional channels before financial operations.
Deepfakes are not just a technological problem. They are a challenge to our ability to navigate reality and trust our perception. The response is not paranoia — but awareness and the habit of checking.
Want to explore your situation?
Talk to an AI psychologist — anonymous, non-judgmental, whenever you need
🌿Start a conversationDisclaimer
This article is for informational and educational purposes only. It is not a substitute for professional psychological advice, diagnosis, or treatment. If you are in a crisis situation, please reach out to a qualified mental health professional or a crisis helpline.