Welcome Citizen!

Sign in to start sharing and discover the best products you can buy today!

Welcome Citizen!

Setup your account or continue reading!

Settings
google-detect-ai-generated-content

Google Developed New Tool To Detect AI-Generated Content

Editor
Content Editor

Celine Low chevron_right

Celine is ProductNation's content editor with a focus on tech social and industry stories. Her previous work includes lifestyle ar ...

There's no denying that generative AI is everywhere and it's here to stay. It has allowed people to create high-quality text, audio, images, and videos with just a few sentences.

However, telling what's real from what's AI-crafted has become increasingly difficult and potentially dangerous with rampant cases of deepfakes and online scams in recent years.

Now, Google offers SynthID, a verification portal that can help with identifying content that's been made with its AI tool.


What Is SynthID?

It is a subtle watermark embedded directly into the AI-generated content. Think of it as a hidden tag marked by invisible ink. The watermark remains detectable even when the content is shared across platforms or undergoes various transformations like compression, resizing, or cropping.

Google's SynthID covers a variety of AI-generated content such as images, videos, texts, and audio files that have been produced using Gemini, Imagen, Lyria, and Veo models.

"Over 10 billion pieces of content have already been watermarked with SynthID," according to Google.


How Does The SynthID Detector Work?

  1. Upload your content
  2. The portal scans in search for the embedded SynthID watermark
  3. The tool identifies the AI-generated content

When a watermark is detected, it highlights the specific portions of the content where the watermark is most likely present. For instance, in an audio file, it will pinpoint particular segments, and for images, it will indicate specific areas that are more likely to contain the watermark.

When content like an artwork is made with Google AI, it's automatically embedded with SynthID. This invisible mark allows Google to identify it as AI-generated, even if someone else tries to claim it as their own.


Though this tool is a welcome development, the skepticism remains: this tool only detects Google's AI-generated content. What about content from other AI models?

Stay updated with ProductNation on here, Instagram & TikTok as well.

News sources: Google Blog

More announcements from Google:

End of Article