Google Launches SynthID Site to Detect AI-Generated Content

Google launched SynthID Detector globally on October 7, 2026, a web tool that allows any user to verify whether an image, video, or audio file was created or edited using AI. The tool scans for invisible SynthID watermarks embedded by Google and its partners, including OpenAI, NVIDIA, Kakao, and soon Apple .

What the Tool Does

- Detection: Upload media at synthid.com to check for SynthID watermarks, which are imperceptible to humans but detectable by algorithms .

- Partner Coverage: Detects AI content from Google, OpenAI, NVIDIA, and Kakao, with Apple support coming .

- Authentication: Requires login via Google, OpenAI, or Apple account .

- Built-in Verification: Also integrated into Google Search, Gemini app, and Chrome, handling over 1 million verification requests daily .

Key Limitations - Not a universal AI detector: Only identifies media watermarked by partner companies using SynthID technology. AI content from non-partner tools (e.g., Meta, Microsoft) will not be flagged .

- Cannot distinguish created vs. edited: A positive result indicates content was “made or edited” with AI, not which specific aspects are synthetic .

- Watermarks can be stripped: Some modifications or processing may remove the watermark, and a negative result does not guarantee content is authentic .

- Daily check limit: Users are limited to approximately 10 checks per day, reportedly to prevent development of watermark-removal tools .