Chapter 14 · Images, sound and video
Deepfakes, watermarks and provenance
Generated media can deceive; watermarks and content credentials help show where media came from.
Chapter 14 · Images, sound and video
Generated media can deceive; watermarks and content credentials help show where media came from.
A generated picture. To your eyes it is just a picture.
A watermark is a clue about where a picture came from. Whether the picture is honest is a separate question.
Pictures, voices and video can now be made that look and sound real. A deepfake is one made to pass as a real person or event. Most generated media is harmless. Some is made to deceive.
Judge five drawings, then see how a hidden mark behaves. Both exercises are about the limits of spotting.
These are drawings made for this lesson. Pretend each one was shared online as a photo. Tap anything that looks off to circle it, then give your verdict. Some pictures have nothing to circle.
A media-literacy toy using cartoon drawings, not real photographs or people. It teaches how to recognise warning signs and why they are not enough. It does not show how to make fakes.
You find no warning signs in a picture. What can you conclude?
Why can a screenshot lose a content credential?
Which habit helps most when you see a shocking picture?