Deepfakes and AI images: keeping a critical eye

Artificial intelligence can generate very realistic faces, voices and videos. A piece of content can therefore be completely made up while looking real. That doesn't make everything suspect, but it changes how you look at what circulates.

A few signs to watch for

Odd details on hands or eyes, an inconsistent background, a voice slightly out of sync with the lips, a source you can't trace: all clues. But the tech keeps improving, so the best reflex is to check the source rather than just the image.

Before sharing a shocking image or video, ask where it comes from and whether reliable sources cover it too. Emotional, instant sharing is exactly what makes false information travel.

When a deepfake targets someone

Fake content targeting a real person can do real harm. Don't relay it, report it, and support the person targeted. If it involves you or someone close, turn to a trusted adult or a support service.