Google is making it easier to remove sexually explicit deepfakes from searches

Google joins a growing number of companies that represent deepfakes for graphic sex.

Characterization made it easier for users to report objectionable images found in search results, including those generated by artificial intelligence tools. While it was previously possible for users to request the removal of these images prior to review, under the new policy whenever such a request is accepted, the company will scan duplicates of the offending image and remove those as well. Google will also try to filter all results that reveal the same search.

“With every new technology development, there are new opportunities to help people – but also new types of abuse we need to fight,” product manager Emma Higham wrote in a blog post. “As the technology of photo-production has continued to improve in recent years, there has been an increase in the production of photos and videos depicting people in sexually graphic situations, being distributed on the web without their consent.”

Google also changed its ranking system, demoting private content that is generally transparent. Even a direct search for obvious deepfakes will bypass the user’s request and instead return “high-quality, potentially revealing content – such as relevant news articles – when available,” the company wrote.

Websites with a high number of pages removed from search under these policies will be demoted in the search algorithm, making it more difficult for anyone to find them. Google says this approach has worked well for other types of malicious content.

Google’s change to the search engine comes just one day after Microsoft asked Congress to create a “deeper anti-fraud system” to combat AI fraud in both images and voice reproduction, and almost a week after Meta’s supervisory board said the social media giant failed its response to a pair of high-resolution, AI-generated images of women in public on its sites.

The US government has taken many steps to curb deepfakes. Recently, the Senate passed a bill that would allow victims of graphic sexual images to sue their creator for damages. And the FCC has banned robocalls with AI-generated voices, which have been on the rise over the past year, especially in the political arena.

They continue to spread, however, and Google has admitted that even with today’s changes to its Search policy, they will continue to appear.

“There is much work to be done to fix this problem, and we will continue to develop new solutions to help people affected by this content,” Higham wrote in a blog post. “And as this challenge goes beyond search engines, we will continue to invest in industry-wide partnerships and professional networking to address it as a community.