Google opens AI detection to everyone

Google launched SynthID Detector to the public on October eighteenth, twenty twenty-six. The tool checks whether images, videos and audio files were created or edited using AI.

How SynthID Detector works

Google's Gemini AI model embeds a watermark in everything it creates. SynthID Detector scans for that watermark by uploading a file to SynthID dot com and waiting for a yes or no result.

It detects more than just Google

SynthID Detector can identify watermarks from OpenAI, NVIDIA and Kakao, in addition to Gemini. Apple participates in SynthID, but support for detecting its creations is still in the works.

Major limitation: it only finds watermarks

SynthID Detector cannot detect AI-generated media from companies that do not use SynthID, such as Meta's tools. It also cannot tell users which parts of a file were AI-generated.

Google limits daily checks to prevent abuse

Users are limited to around ten checks per day to prevent them from using the detector's output to find ways to remove SynthID watermarks. Users must be logged into a Google, OpenAI or Apple account.

How it compares to other detectors

OpenAI runs its own detector. Meta's Content Seal detector sometimes failed to spot watermarks, according to a two thousand twenty-six Reuters report. No reliable detector exists that spots generative AI across the board.

“a 2026 report from Reuters found that it sometimes failed to accurately spot them”

The broader problem remains unsolved

Many researchers have developed AI-spotting tools but they are not available to the public. General purpose AI detectors that are openly available have unreliable results and often flag human-created work as AI content.