Google has updated its policies to tackle explicit fake images in search results.
The rise of sexually explicit deepfakes, which disproportionately affect women, has become a significant issue. Last week, a bipartisan bill to combat the distribution of such intrusive content passed unanimously in the Senate. In response, Google is stepping up its efforts to combat this harmful material. The company outlined its strategy on its blog, detailing a two-pronged approach to prevent deepfakes from appearing in search results.
If you’ve been a victim of an explicit deepfake or revenge porn, your immediate concern is likely to get the offending content removed. Historically, this has been a challenging process. Google has now updated its system to make content removal more manageable. According to Google, once the offending content is removed, the search engine will try to filter out all related explicit content and remove any duplicate posts.
To have an explicit post removed from Google Search, the following criteria must be met:
- You are clearly identifiable in the image.
- The image is fake and falsely depicts you in a nude or sexually explicit situation.
- The image was distributed without your consent.
Google has also updated its ranking system. The first step is to create a ranking that minimizes the surfacing of explicit content. For search queries with a high rate of inappropriate results, the search engine will attempt to show non-explicit results, such as news articles. Google claims that these updates have reduced deepfake entries by over 70%, though real-life data on this statistic would be helpful.
Google acknowledges the ongoing challenge of determining whether content is consensual, such as an actor’s nude scene versus a deepfake. The company states, “We’re making ongoing improvements to better surface legitimate content and downrank explicit fake content.”
Additionally, Google will penalize sites with too many flags regarding content removals. If a page has numerous explicit content removals, Google will decide when and if the site will appear in search results.
While these are relatively small steps in the broader context, they provide victims of deepfake and revenge porn with tools to fight back. If Google and other AI companies can establish rules and regulations to address these issues, it may reduce the need for government intervention.
