Artificial intelligence synthesis models are being used to create fake images, which are circulating on the internet. Because of this, it has become difficult to differentiate AI-generated images from the real ones. In a blog post, Google mentioned that a 2022 study conducted by Poynter showed that 62 per cent of internet users are of the opinion that they come across fake news and misinformation on a daily or weekly basis.
At Google I/O 2023, Google announced new features that will help users detect AI-generated fake images in search results. These new features will seek to identify the origins of an image and label other AI-generated images in search results in an effort to curb the spread of disinformation.
One of these new tools, called ‘About this image’, will assist internet users in quickly assessing the credibility of images. With 'About This Image' feature, users will be able to click three dots on an image in Google Images results, search with a picture or screenshot in Google Lens, or swipe up in the Google app to learn more about an image's history, indexing and first appearance.
This new feature will be rolled out in the coming months. Initially, it will be available exclusively to users in the US and will only support English. Google says users will also be able to access this tool by right-clicking or long-pressing on an image in Chrome on desktop and mobile later this year. This context about an image will help a user to decide on its reliability.
Google also plans to label the images generated by its AI tools with special markup that will clearly indicate its AI origins. Such markups include information in images metadata and watermarks.
The tech giant also said it will work with other platforms to add similar labels to AI-generated images. Midjourney and Shutterstock have agreed to sign on to the initiative.