Google Eradicates Deepfake By Creating Watermark

Director of Government Relations and Public Policy of Google Indonesia Putri Alam when discussing how to eradicate disinformation (photo: Rana/VOI)

JAKARTA Technological developments have created new problems for Google after Deepfake, engineering photos or videos with Artificial Intelligence (AI), mushroomed.

Deepfake itself can be a serious problem if left alone because some parties can use this tool to bring down certain parties and spread disinformation.

To eradicate this sophisticated tool, Google Indonesia stated that their company has created a watermark to mark images created by AI, especiallydeepfakes.

"However, before that we were clear, once again, our community guidelines that we prohibit content that personalizes someone, especially if the goal is false or fake," said Director of Public Relations and Policy of Google Indonesia Putri Alam at the #YukPahamiPemilu event on Wednesday, September 20.

When Google finds photos usingdeepfake that are deliberately disseminated by irresponsible parties, Google will immediately withdraw the upload or in other words take down.

In line with Putri's statement, the Google DeepMind Team has released a tool called Synthedite which is designed to mark images or provide watermarks on images produced by AI.

This tool was released at the end of last August with performance that is not capable of being performed by the human eye.Watermarkini embeds in image pixels and will make AI images easy to detect by DeepMind.