In the middle of rage between the U.S. Department of Justice and Google over “illegal monopoly”, Google made a big announcement. Google announced that it clearly going to label all AI images and AI-edited images for better search results and to assist visitors to understand what they are consuming. Yes, you read it right, Google labels AI images that appear to search and search ads.
Though merely labeling AI images is not sufficient for a better search, experts think it is an admirable initiative. In the near future, we may also find AI-labeled content which is a major concern among publishers and tech companies.
This decision comes after Google’s recent dialogues between Google and the Coalition for Content Provenance and Authenticity (C2PA). C2PA is an alliance between Adobe, Arm, Intel, Microsoft, and Truepic. The major goal of C2PA is to address the prevalence of misleading information and media content by developing content credential standards.
Once fully rolled out, Google integrates this feature into its all products including search, images, and lens. It helps users to understand the AI-generated content, how it developed, when it was edited, etc. Users can access these labels directly from the search engine under the “About this page” section.
The partnership with C2PA may help Google to the establishment of a new era of standards for AI-generated concern. All the related information is shared by Google through its blog, “The Keyword”. You can read more about this announcement by clicking here. Along with this, Google also signals to advance its watermarking tool SynthID, created with DeeMind for better transparency with AI-generated text, images, audio, and video.
How Google AI Labels to Images Help Users
AI-generated content is a major concern for users as they are unable to identify an image if it is created by a human or generated by an AI. However, labeling AI images is not enough to bring transparency into the internet as we have seen recently deep fake videos creating controversies. In a recent incident, a UK engineering firm “Arup” fell victim to a £20m deep fake scam, according to the Guardian post.
This is not the only case, we have seen heated arguments between Joe Biden and Donald Trump during presidential election speeches. Google is ready to work with C2PA technology to label AI content in the search and other associated products. With this initiative, we are assuming that Google soon come with the same label for text, audio, and video content.
By easy identification of original content and AI-generated content, users get help on what to consume and what not to consume. The tech giant will use metadata to label the images across search, images, and lenses to identify the image’s origin. In its blog, Google clearly declared that it is going to use this feature not only to search but also target to ads.
Also read: Securing Apps: How to hide apps on Google Pixel phones with Private Space
How users can identify if an image is AI-generated?
To improve the search and bring more transparency for users, Google has not only partnered with C2PA but also partnered with other organizations and governments. Frontier model forum, MLCommons, C2PA, G7 code of conduct, UK & US AI safety institutions, etc.
These partnerships greatly improve the lost reputation of Google that we have seen since the AI’s inception in the search. When there is transparency and clear labeling, users can identify images that can use their discretion for the truth behind the content they consume. It also helps in the prevention of rumors and fake information everywhere and deep fake is the biggest example of this.
Google has also mentioned how a user can identify if an image is AI-generated or humanely created. You can follow the below steps to find out this yourself. However, you may not be able to see the AI label as it may not be available in your region yet.
- Open the Google app or Chrome browser on your device.
- Search for any image in the search engine and click on the “Image” tab in the toolbar.
- Now, click over any image that you want to know about.
- It will open in a small carousel at the top-right side.
- Click on the three-dot button top-right corner of the image (as shown in the above screenshot)
- Click on the last option “About this image” to learn more about the image.
- It will lead you to a new tab and you can see all information regarding the image. This page majorly tells about when the image was first time appears on Google and where it is found. Now, it will also display how an image is generated.
- However, you may not be able to see the “AI generated label” if it is not available to your region as it is still rolling out.
- Similarly, you can check any image available on the internet. We are also hoping for this type of labeling in all the content we consume.
Conclusion
With the inception of AI and AI-generated content, users and publishers have been demanding for a long to label AI-generated content. Google labels AI images are currently the major buzz among publishers. An announcement of labeling AI images is considered a positive step towards improving the search and bringing transparency.
Google has partnered with C2PA, a major organization known for the prevention of misleading information by developing technical standards. We are expecting and optimistic that the step of labeling will bring a positive impact to the search.