Headlines:

• Google finally unveils its SynthID tool that helps users detect AI-generated images.
• The tool does so by imprinting invisible watermarks on AI-generated images, which can only be seen using this tool.
• Google has launched this tool to curb the spread of AI-generated “deepfake” images and promote non-disruptive use of AI-generated content.

With the advent of AI-generated content booming and on a constant rise, many bloggers, artists, photographers, video editors, and even musicians are getting afraid of its development. Therefore, to make matters easier for everyone, letting them differentiate between realistic AI-generated images and real ones, Google has blessed us with SynthID – an AI-generated picture detector.

While this tool is only available to select Google Cloud users for now, it will soon be made available to all Google users. Currently, Google is testing its prowess and effectiveness by using it alongside Imagen, Google’s tet-to-image model.

How Does SynthID Work?

Google’s SynthID makes it easier than before to detect AI-generated images by using digital watermarks. 

This tool uses two highly advanced and trained AI models. One of the models is used for detecting AI-generated images, while the other one is meant for adding imperceptible digital watermarks and identifying them.

For this AI model to identify an AI-generated image, it scans images in three stages (called Confidence Levels). If an image passes all three levels, it will then get digitally watermarked by this tool, which can only be seen using this tool.

Important early progress on watermarking from @GoogleDeepMind with beta release of #SynthID https://t.co/FysmSerkht

— Sundar Pichai (@sundarpichai) August 29, 2023

Google has claimed that even if an AI-generated image gets manually edited, cropped, etc., later on, it still leaves signs behind that it’s an AI-generated image. Therefore, even heavily edited AI-generated images will still get detected by this tool!

How (And Whom) Will This Tool Help?

SynthID will be a great boon (and crutch) for marketers as they will be less invested in using AI-generated images for their marketing campaigns. This is expected to build proper attribution, which will help marketers and brands build trust with customers.

More Resources:

• Meta’s Llama 2 Finally Enters The AI Chatbot Market
• SurferAI vs. Autoblogging vs. Jarvis AI
• LinkedIn Can Now Catch Fake Profiles Using Its New AI Image Hunter