Google unveils SynthID: A step forward in combating misinformation

Google introduces SynthID, a new watermarking method for AI-generated images, aiming to curb the spread of fake news.

Google has taken a new initiative in the fight against misinformation by launching SynthID, a technology that adds an invisible watermark to AI-generated images. What does this mean for the digital landscape?

How SynthID works

With SynthID, Google can embed a permanent watermark in images generated by Imagen, one of their latest text-to-image generators. This watermark remains intact even with subsequent edits.

Testing and accuracy

Google has confirmed that the technology has proven to be effective against many types of image manipulations in their internal tests, although it is not foolproof.

Future prospects of SynthID

SynthID is currently in the beta phase and available to selected Vertex AI customers. Google has plans to further develop the technology.

EU’s perspective

Vera Jourova, Vice President of the European Commission, has urged major tech companies to take steps to identify and label manipulated content.

Other players in the market

Besides Google, other companies like Truepic and Reality Defender are also working on technologies to identify manipulated content.

Challenges

Google has taken an important step in the direction of combating misinformation with SynthID, but there are still challenges to be overcome.

Did You Know?

• SynthID was created in collaboration between Google’s DeepMind and Google Cloud.
• AI-generated images have previously led to controversies, such as images of public figures.
• Google also has another tool, “About this image,” that provides additional information about images.