Forensics test spots AI media using simple geometry checks

This gives video teams and trust-and-safety workflows a direct visual check for AI-generated media, rather than relying only on automated detection. For streaming, that matters anywhere authenticity review touches UGC, moderation, or newsroom-style verification. The article does not describe the test’s accuracy, false-positive rate, or implementation cost, so those limits remain unknown. The next concrete signal to watch is whether the method is documented in more detail beyond this report, especially any examples showing how the geometry test is applied to images versus video.