Google’s Veo 3, released in May 2025, marks a leap in AI video generation—delivering eight‑second clips with synchronized visuals and realistic audio that are easily mistaken for real footage. Al Jazeera and TIME have demonstrated how it can fabricate protest scenes, news reports, and violent events with minimal prompting. Although Google embeds both visible watermarks and invisible SynthID markers and applies filtering mechanisms, critics argue these safeguards are insufficient, as manipulated videos—ranging from false election fraud to hateful content—are already circulating widely