Google Tingkatkan Tools Deteksi AI dengan SynthID
Peningkatan dan Perluasan Deteksi
Google announced its upgrade of the SynthID platform with the addition of the SynthID Detector tool, which now supports media detection beyond Google's own models. Google officially expanded the reach of its AI watermarking system to include images, videos, and audio files generated by artificial intelligence from other major technology companies, including OpenAI, NVIDIA, and Kakao. Furthermore, the company confirmed that AI-generated media from Apple will soon be added to the list of supported sources as well. SynthID remains the primary tool developed by Google to identify AI-created images, videos, and audio content across all supported platforms.
Cara Kerja dan Pencapaian Google
Google's AI models embed a specially designed watermark that is completely invisible to the human eye directly into AI-generated images. The SynthID system is built to detect these embedded watermarks and inform users when they are viewing media that has been generated or manipulated by artificial intelligence. According to Google, the company has successfully watermarked more than 180 billion images and videos since the tool was first launched in 2023. In addition to visual media, Google has also embedded watermarks into 240000 years worth of audio content over the same period.
Cara Menggunakan SynthID Detector
To use the SynthID Detector tool, users simply need to visit synthid.com and log in using one of their Google, Apple, or ChatGPT accounts. After logging in, users can upload any supported media file and allow the tool to analyze the content automatically. The SynthID Detector will then display whether it has detected a SynthID watermark and whether the uploaded content is designated as AI-generated or AI-manipulated.
Keterbatasan Alat Deteksi
It is important to understand that the SynthID Detector is not a general AI detection tool in the traditional sense. The system specifically detects only the watermark that certain AI models embed into their generated media outputs. If you upload AI-generated content that was produced by a model which does not embed the SynthID watermark, the SynthID Detector will not be able to identify that the media is AI-generated. This means the tool has limited scope and cannot verify AI origin for all types of synthetic media currently in circulation.