ADVERTISEMENT

Google is working on ways to tell you if an image is AI generated

The Telegraph tells you how...

Mathures Paul Published 16.05.23, 08:11 AM
An AI-generated image of Pope Francis in a puffer jacket

An AI-generated image of Pope Francis in a puffer jacket Sourced by the correspondent

AI-generated images can obviously be used for misinformation. Agreed, it’s tough to believe one’s eyes when Pope Francis is seen wearing a puffer jacket. But what about former US president Donald Trump being arrested? Google is taking steps towards fighting misinformation arising from AI-generated images.

The company is introducing two new features to Google Search to help understand the content and context of the image that’s before the user.

ADVERTISEMENT
An AI-generated image of former US president Donald Trump getting arrested

An AI-generated image of former US president Donald Trump getting arrested

The first new addition is going to hit Google Search in the “coming months” and will introduce more context about images in an “About this image” feature. It will show information about when an image first appeared online and where else the image has been featured (like media websites or fact-checking sites). This will help one understand if an image is being used in a misinformation campaign.

According to a 2022 Poynter study, 62 per cent of people believe they now encounter misinformation daily or weekly. “About this image” feature will help tackle this. To use the feature, users can click on the popular three dots on an image in the Google Images search results. Users can also search with an image.

The only issue is that the feature will initially be limited to the US and only in English.

The second option involves creating generative images. Google’s own AI-generated images have a markup in the original file to give context if the image is accessed outside of Google’s platforms. The company has said other creators and publishers will be able to label their images using the same tech.

Google is bringing AI art generation to Bard with the help of Adobe. Adobe has announced that Firefly, its collection of AI models that can be used to generate media content, is coming to Bard alongside Adobe’s free graphic design tool, Adobe Express. Firefly will also be an important tool to identify AI art. There will be tech to trace breadcrumbs back to the original artist it’s referencing. The partnership comes soon after the launch of Image Creator, Microsoft’s generative AI tool built into its Bing Chat chat bot.

Follow us on:
ADVERTISEMENT