Google Takes Action Against Nonconsensual Deepfakes with Ranking System Update

In an effort to combat the growing issue of nonconsensual deepfake content, Google has announced a major update to its ranking systems, designed to reduce the visibility of explicit fake imagery in search results. The new measures, outlined in a blog post by Google product manager Emma Higham, aim to provide users with more responsible content, particularly when searching for deepfakes involving specific individuals.

What’s Changing in Google’s Ranking System?

The updated ranking system targets explicit deepfake content by downranking it in search results. When users search for deepfakes that involve nonconsensual content, the system will now prioritize "high-quality, non-explicit content," such as news articles, educational resources, or discussions about the social implications of deepfakes. Higham emphasized that these changes aim to steer users away from harmful material and instead help them access information on the societal impact of deepfakes.

The update has already made a significant impact, with Google reporting a 70% decrease in the visibility of explicit deepfake images in search results. This move is part of Google's broader effort to address the negative consequences of deepfakes, which have been used to exploit individuals without their consent.

Challenges in Identifying Fake Content

One of the challenges in curbing deepfake content is distinguishing between consensual and nonconsensual material. For example, explicit scenes featuring actors in films may appear similar to deepfakes but are consensually produced. Higham noted that while this distinction can be challenging, Google’s system is designed to prioritize content that aligns with ethical standards.

To improve the identification of nonconsensual deepfakes, Google has incorporated a new factor into its ranking systems: whether a site has removed pages under Google’s policies for fake explicit content. Websites that have a history of violating these policies will be penalized and pushed further down in search results.

Simplifying the Removal Process for Victims

Another important update is the improvement of Google’s deepfake removal process. Victims of nonconsensual deepfakes will now find it easier to request that such content be removed from search results. Additionally, once a deepfake is successfully removed, Google's systems will aim to filter out all related results, including duplicate versions of the content, and prevent similar searches from returning harmful imagery.

More Work Ahead

While the new updates represent an important step in tackling the issue of nonconsensual deepfakes, Higham acknowledged that Google still has "more work to do." The company will continue to enhance its systems and develop new solutions to better support victims and reduce the spread of harmful deepfake content across the web.

This move by Google follows a call from the White House two months ago for tech companies to take more responsibility in addressing the growing threat of explicit deepfake imagery. With these updates, Google is making strides to ensure that the search engine promotes a safer, more responsible internet for all users.