Google modifies search engine to aid hide obvious deepfakes

Google is rolling out recent online safety features that make it easier to remove obvious deepfakes from Search at scale and prevent them from appearing high in search results.

When users successfully request removal If explicit, non-consensual imitation content featuring them appears in search results, Google’s systems will now also attempt to filter out any explicit similar search results about them and remove any duplicate images.

“These safeguards have already been proven effective against other types of unwanted images, and now we’ve made it possible to do the same for fake sexual images,” said Emma Higham, Google product manager it was said in the announcement“These actions are intended to provide people with additional peace of mind, especially if they are concerned about similar content appearing about them in the future.”

Google’s search rankings are also being adjusted to better handle queries that carry a higher risk of displaying overtly false content. For example, searches that intentionally look for deepfakes depicting a real person (such as the sexually explicit AI-generated images of Taylor Swift that were made public earlier this year) should instead display “high-quality, non-classified content,” such as relevant news articles. Sites that receive a significant number of takedowns for overtly false images will be demoted in Google’s search rankings.

Google says previous updates have reduced the exposure of explicit image results for queries specifically looking for deepfake content by more than 70 percent this year. The company is also working on a way to distinguish between real explicit content — such as scenes of an actor getting consensually nude — and explicit imitation content, so that it can still show real images while downgrading deepfakes.