Image: Google.
Generative artificial intelligence, particularly through image generators, is capable of producing incredible images. The viral image of Pope Francis wearing a white jacket, for example, fooled the entire internet.
As this example shows, this technology poses a high risk of spreading false information. This is why Google is launching a new tool to help its users distinguish fact from fiction.
During the Google I/O conference, the company announced the creation of the “About this image” functionality in its search engine. It aims to help users determine whether an image is real by providing them with crucial information about its origin, among other things.
Metadata, source and recovery
Last Wednesday, Google finally unveiled its “About this image” feature: it will display, with a simple click, the history of the image, how other sites use and describe it, and above all, its metadata.
An image’s metadata is valuable because it contains information where one can find details that the creator or editors of the image themselves added to the image, ensuring that they are properly credited for their work.
Additionally, many AI image generators – such as those from Google or Adobe – indicate in the metadata that the image was generated by AI. Therefore, easier access to metadata will allow users to quickly access a clear marker informing them that the photo is fake.
To access the “About this image” functionality, simply click on the three dots appearing next to an image in the Google Images results, or click on About this result > More information on this page in search results.
Fact checking
Google is also introducing the Fact Check Explorer tool, which provides journalists with a platform to easily obtain more information about an image or topic.
To use this tool, all the user needs to do is upload an image or the URL of what they are looking to check. The tool will then check if the image has already been verified and present this information.
According to Google, since the feature’s launch, more than 70% of beta users reported that the new image features helped reduce their investigation time.
At the request of testers, Google also revealed the beta version of a FactCheck Search API that users can integrate with their own fact-checking solutions.
Authenticate images as soon as they are shot
Google isn’t the only company trying to curb misinformation from AI-generated images. Camera maker Leica has launched the world’s first camera to feature content references.
This means that when the photo is taken, the photo automatically includes metadata including details like who took the image, and when and how it was taken.
Source: ZDNet.com