Recognising disinformation and deepfakes

In short: Disinformation refers to deliberately spread false information, while deepfakes are deceptively realistic image, audio, or video forgeries created with artificial intelligence. Both phenomena pose growing challenges for media literacy.

In more detail: Deepfakes are created using neural networks that learn to convincingly imitate faces, voices, or movements.

In detail

• Source checking: Verifying who originally published a piece of information and whether the source is trustworthy is among the most important steps.
• Visual inconsistencies: Unnatural blinking, odd shadows, blurry transitions at facial edges, or inconsistent lighting can indicate deepfakes.
• Audio clues: Unnatural emphasis, a metallic sound, or missing background noise sometimes point to synthetically generated voices.
• Checking context: A reverse image search and tracing where content was originally published often reveals manipulation.
• Emotional manipulation: Content designed to trigger especially strong emotions like anger or fear requires heightened scepticism.
• Fact-check portals: Independent organisations systematically check viral claims and videos and publish their findings publicly.
• Technical detection tools: Specialised software can detect subtle digital artefacts that often go unnoticed by the human eye (clickbait-und-schlagzeilen-durchschauen, online-bewertungen-richtig-einschaetzen).