Overview

• Since the June 13 escalation, social platforms have been flooded with AI-generated videos depicting bombed cities and downed jets that never existed.
• Investigators at India Today and Full Fact traced viral clips to Google’s Veo 3 tool by spotting faint ‘Veo’ watermarks and confirming their origin with SynthID.
• Both Iranian and Israeli officials have shared fabricated footage on their official channels to advance their respective narratives of the conflict.
• Detection remains challenging for casual users because invisible SynthID tags require downloading and uploading videos to a separate verification site.
• Google now embeds invisible SynthID watermarks in all AI output and has added a visible watermark to Veo 3 videos to help platforms and fact-checkers identify fakes.