Google to Demote Websites Harboring Deepfakes

The rapid advancement of AI technology has made it easier for individuals to create non-consensual sexually explicit deepfakes, prompting Google to update its search engine algorithm and removal request process to combat this growing problem.

Key Changes to Google’s Approach: Google is taking proactive measures to address the issue of unwanted AI deepfakes:

• The removal request process for victims will be streamlined, making it easier to report and remove such content.
• When reported AI deepfakes are identified, Google Search will automatically filter out related search results that might appear in the future, reducing the need for repeated reporting of similar or duplicate images.
• Google’s search algorithm will be updated to better tackle the issue of non-consensual AI deepfakes.

Demoting Offending Websites: Google will take action against sites that repeatedly host non-consensual AI deepfakes:

• Websites that consistently harbor such content will be demoted in Google Search rankings.
• This approach has proven effective for other types of harmful content, and Google’s testing suggests it will be a valuable way to reduce the appearance of fake explicit content in search results.

Technical Challenges and Effectiveness: While Google aims to differentiate between consensual and non-consensual sexually explicit content, it acknowledges the difficulties involved:

• Distinguishing between real sexually explicit content made consensually and AI-generated media made without consent is a technical challenge.
• Despite these challenges, Google claims that the changes already implemented have reduced the resurfacing of such deepfakes by more than 70%.

Legislative Efforts to Protect Victims: US officials are pushing for laws to protect victims of non-consensual deepfakes and hold platforms accountable:

• Senator Ted Cruz (R-Texas) proposed the Take It Down Act, which would criminalize the publication of non-consensual sexual deepfakes and require social media platforms to remove them.
• The Senate passed the Defiance Act, allowing victims to sue those who have created or shared unwanted sexual deepfakes.

Broader Implications: The issue of AI deepfakes extends beyond sexually explicit content, with potential consequences for politics and society:

• AI deepfakes have been used to discourage primary voting through robocalls impersonating President Joe Biden.
• Fake videos of Vice President Kamala Harris have gone viral on social media platforms like TikTok and Twitter/X, raising concerns about the spread of misinformation and the potential violation of platform rules.

As AI technology continues to advance, it is crucial for tech companies, legislators, and society as a whole to address the challenges posed by non-consensual deepfakes and develop effective strategies to protect victims and maintain the integrity of online information.