AI Tool of the Week: this tool checks if an image, video or audio is AI-made

Summary

Google’s SynthID Detector checks images, videos and audio for invisible AI watermarks, giving professionals an initial verification step while highlighting why a missing watermark does not prove content is authentic.

The AI capability unlocked today is Google’s SynthID Detector.

What problem does it solve?

A photo arrives with a damage claim. A voice note that sounds like your CEO asks for urgent payment. You suspect AI, but cannot check, so you either trust it or lose a day escalating it.

Finance, HR and legal teams increasingly face the challenge of assessing potentially synthetic media. Google's SynthID Detector offers a first check: upload a file to see whether it carries a SynthID watermark, an invisible marker embedded by compatible AI systems. The technology is designed to help identify AI-generated or AI-edited content, though its coverage depends on whether the file carries a detectable watermark.

What can it do?

• Verify before acting: Check a suspicious file in minutes instead of escalating it.
• Cover every format: Test 11 image, 3 video, and 6 audio formats, from HEIC to MP3.
• Resist typical edits: Google says watermarks are built to survive compression, cropping, and filtering.

Example

A finance manager receives a vendor’s bank-change request with a supporting voice note, and payment is due by evening.

• Sign in: Go to synthid.com and sign in.
• Upload the file: Choose “Upload file” and select the voice note, for example an M4A or WAV.
• Read the result: The site reports “watermark detected” or “watermark not detected.”
• Interpret carefully: Detected means it was likely generated or edited by a SynthID-enabled model; not detected rules nothing out.
• Gather more evidence: Call the vendor on a known number and escalate to security before paying.

Edge case worth knowing: It cannot tell AI-created from AI-edited media, and rare false triggers can occur.

Jaspreet Bindra