Deepfakes are videos, images or audio recordings manipulated by artificial intelligence to give the illusion that a person has said or done something they never said or did. The term combines ‘deep learning’ and ‘fake’, referring to fictitious content made extremely credible by artificial intelligence.
While some deepfakes are used for fun, others are used to deceive, manipulate public opinion or commit fraud.
How are they used?
Deepfakes are exploited in several contexts:
- Disinformation and manipulation: Fake videos of public figures can be used to spread lies.
- Fraud and scams: Scammers imitate the voice or face of a relative or superior to extort money.
- Fake adverts: Fake videos show celebrities touting profitable investments that are in fact fictitious.
- Damage to reputation and cyberbullying: Some deepfakes are created to damage a person's reputation by attributing words or actions to them that they have never committed.
How can you detect a deepfake?
Spotting a deepfake can be difficult, but there are some clues that should alert you:
- Strange facial expressions: Look to see if the facial movements seem natural. An unusual blink or a lack of blinking can betray a fake video.
- Inconsistencies in the audio: The voice may seem slightly robotic or out of sync with the lip movements.
- Blurry or unstable background: Deepfakes sometimes have blurry backgrounds or objects that disappear or become distorted. Inconsistent lighting can also indicate manipulation.
- Watch the videos in slow motion: at a slow playback speed, errors in lip movements, facial expressions or background are more visible.
- Do a reverse image search on Google Images to see if the video or photo exists elsewhere.
How can you protect yourself?
- Remain critical: Don't immediately believe everything you see or hear online.
- Always check the sources: Rely on recognised media and official news sites before sharing a video or information.
- Analyse the context and timing: check the context in which the content was shared and the associated dates, as deepfakes can manipulate the chronology of events.
- Raise awareness among those around you: explain to your friends and family what a deepfake is and how to recognise them.
- Secure your personal data: protect your photos and videos by limiting access to them on social networks.
- Be careful with artificial intelligence tools: some applications request access to your voice or photos. Only authorise this if you really trust the tool.
- Report suspicious content: see something suspicious on social media? Report it (using the three dots at the top right of the post). This will help you to protect others.
3
