“Google’s Bold Stand: Shielding Users from AI Deception”
Google is actively working to tackle the rise of AI-generated content, like deepfakes, by teaming up with various industries to identify and watermark synthetic online content. Markham Erickson, VP of Government Affairs at Google, shared this plan during an interview at the Global Technology Summit.
This effort is a response to the growing worries about the potential threats posed by synthetic media. Erickson highlighted the importance of industry-wide discussions to create consistency in transparency and awareness for users encountering AI-generated content on different platforms.
In November, YouTube, a Google subsidiary, took a proactive step by labeling all content created using YouTube’s generative AI products as altered or synthetic. This labeling also applies to synthetic content generated outside of YouTube, thanks to participation in a cross-industry collaboration.
Dealing with content from non-participating entities presents challenges, and Erickson acknowledged ongoing efforts to develop technology for identifying synthetically created content from unwilling sources.
As part of Google’s commitment to combat political misinformation, they’re working on classifiers to identify AI-generated synthetic media. Google has achieved an impressive 99% accuracy in detecting AI-generated audio, demonstrating the effectiveness of AI in addressing evolving challenges.
Detecting synthetic media, including deepfakes, has become more complicated due to advancements in AI art programs and photo editing features. Google’s policies now require advertisers to disclose digitally altered or generated content in election ads on Google and YouTube.
YouTube has also set disclosure requirements for all content creators, especially in sensitive areas. Violating these terms can result in consequences such as content removal and suspension from revenue-generating programs.
Acknowledging ongoing challenges, Erickson stressed the importance of effective policy execution. The article highlights YouTube’s commitment to staying ahead of those trying to manipulate the system, using AI as a crucial tool for model training and content moderation.
