Google introduces new features to help identify AI images in Search and elsewhere

Google is taking its first steps to fight the potential for misinformation from AI-generated photographs. At its I/O developer occasion this afternoon, the corporate introduced two new options coming to Google Search that goal to assist folks higher perceive the content material and context of the picture they’re viewing in search outcomes. This contains including extra background details about the picture itself to the “About this picture” function, in addition to new markup that can enable tagging photographs as “Synthetic Intelligence generated” within the file itself.

The corporate says the brand new options are an extension of its current work, which is able to embrace larger visible literacy and assist folks extra rapidly assess whether or not a picture is dependable or artificially produced. Nonetheless, these instruments alone will seemingly not deal with the broader downside of AI imagery getting used to mislead or misinform – most of which is able to occur outdoors of Google’s partitions and the place creators do not play by the principles.

The primary new add-on, coming “within the coming months”, will supply extra context about photographs within the “About this picture” function.

Particularly, it’ll embrace data reminiscent of when photographs and related photographs have been first listed by Google, the place the picture first appeared on-line, and the place else the picture was seen on-line. The latter may embrace issues like information media web sites or fact-checking websites that might doubtlessly lead internet searchers to be taught extra in regards to the picture in query – together with the way it might need been utilized in misinformation campaigns.

Based on the 2022 Poynter research, Google states that 62% of individuals now imagine they encounter each day or weekly misinformation – an issue Google hopes to resolve with the “About this picture” function.

Whereas Google has been proactive in selling the software earlier than it’s utterly inundated with synthetic intelligence-generated photographs that aren’t instantly apparent at first look – for instance Pope Francis in the hedgehog jacket or Trump and Biden as best friends – additionally one other instance of how AI applied sciences are launched into the wild earlier than strong methods are deployed to mitigate the potential injury they will do.

Google explains that to make use of “About this picture”, customers should click on the three dots on a picture in Google Pictures search outcomes, search with a picture or screenshot in Google Lens, or swipe up within the Google App. You might be on a web page with a picture you wish to be taught extra about.

Later this 12 months, customers will have the ability to entry the function by right-clicking and long-pressing a picture within the Google Chrome internet browser on cellular and desktop.

Nonetheless, the function will initially be restricted to the US and can solely be obtainable in English.

In the meantime, as Google rolls out its personal generative picture capabilities, a second new function will make sure that all of Google’s AI-generated photographs have a markup within the authentic file that provides customers context if the picture is accessed from outdoors of Google platforms.

Picture Sources: Google

Creators and publishers will have the ability to add related markups to their very own AI-generated photographs. Doing so will add a tag to pictures in Google Search outcomes that can mark them as Synthetic Intelligence Generated.

Google says a number of publishers are poised to undertake the function, together with Midjourney, Shutterstock, and others.

Picture Sources: Google

Google says it’ll even be obtainable “within the coming months,” however did not share a extra particular timeframe.