Google’s New Measures Against Deepfakes in Search Results
With constant improvement and the launch of new technologies, today people are able to perform multiple tasks in less time.
No doubt, the technology advancements are helping us in different ways, but at the same time, Google is even encountering new forms of abuse.
For example, with the development of generative imagery technology in recent years, Google has encountered various cases in which people are producing images and videos of people in sexually explicit contexts and sharing them across the web without their approval.
We all remember Rashmika Mandanna’s Deepfake case. Similarly, more explicit “deepfakes” were tracked by search engines that could cause extreme distress to those who are impacted by them.
Analyzing the data, Google decided to invest in long-standing policies and systems that would provide more control over such content.
Finally, in the recent updates, Google has taken some necessary steps to reduce deepfake content visibility over search results.
Key Components Of The Update
Manual Removal Process
Google has expedited its content removal procedure in order to assist those who have been impacted by deepfakes. Google’s systems will filter similar explicit results for relevant queries once a request to delete non-consensual explicit fake content is authorized. In addition, Google will run scans over its search results for duplicates and eliminate them if an image/video is deleted.
Ranking System Updates
Google has made enhancements to its ranking algorithms to guarantee that high-quality and non-explicit information takes precedence over deepfakes, particularly for queries containing the names of individuals. For example, news stories are more likely to show up in search results than deepfake material. The exposure of explicit fake images for these queries has been successfully decreased by over 70% thanks to these modifications.
In fact, websites with fake explicit imagery content might see a difference in their overall search rankings.
Penalizing Low-Quality Sites
Google is also examining patterns among different websites. A website will be considered low quality by Google if a large number of its pages have been eliminated because they contain explicit fake content. As a result, Google devalues these websites in search results, a tactic that has been successful in lowering offensive content.
Industry Context
Google also initially faced difficulty distinguishing between explicit fake content produced by deepfakes and actual consented explicit content, such as an actor’s nude scenes. Google is constantly enhancing its algorithms to more accurately detect and rank fraudulent explicit content while endorsing authentic stuff.
In fact, Google is collaborating with industry partners and experts to tackle this issue more comprehensively.
Emma Higham, a Product Manager at Google, emphasized that these updates are significant.
Google will continue to develop new solutions and work with industry partners to protect individuals from explicit fake content.
With these modifications, Google is putting more of an emphasis on content quality and user safety as it adjusts its search engine to the changing digital environment.
Stay Tuned for more such information.
