Another promising security tool that Google introduced at the I/O event.
At the last Google I/O conference for developers, the head of DeepMind, Demis Hassabis, presented many new tools based on artificial intelligence, including the Veo video generator, search engine updates and the upgraded Gemini model. However, special emphasis was placed on the updated SynthID digital labeling system, which is now able to recognize and label audio, video and text materials created using AI technologies.
The need to track such content becomes more and more acute as generative AI develops, especially in cases of malicious use. There are already many examples of neural networks being used to spread deepfakes, including political disinformation and sexualized content.
The SynthID system was announced in August last year as a tool that can be used to embed water signs that are invisible to the human eye and can be recognized by special algorithms. This approach differs from other proposed digital labeling standards, such as C2PA, which provides for the use of cryptographic metadata.
In the fall of 2023, Google also integrated the digital tagging feature for audio content into SynthID-it allows you to embed inaudible tags in music tracks created using DeepMind's Lyria model. SynthID is one of several measures being developed to ensure the security of AI technologies and prevent their abuse. Even President Biden's administration recently instructed federal agencies to develop guidelines in this area.
At the last Google I/O conference for developers, the head of DeepMind, Demis Hassabis, presented many new tools based on artificial intelligence, including the Veo video generator, search engine updates and the upgraded Gemini model. However, special emphasis was placed on the updated SynthID digital labeling system, which is now able to recognize and label audio, video and text materials created using AI technologies.
The need to track such content becomes more and more acute as generative AI develops, especially in cases of malicious use. There are already many examples of neural networks being used to spread deepfakes, including political disinformation and sexualized content.
The SynthID system was announced in August last year as a tool that can be used to embed water signs that are invisible to the human eye and can be recognized by special algorithms. This approach differs from other proposed digital labeling standards, such as C2PA, which provides for the use of cryptographic metadata.
In the fall of 2023, Google also integrated the digital tagging feature for audio content into SynthID-it allows you to embed inaudible tags in music tracks created using DeepMind's Lyria model. SynthID is one of several measures being developed to ensure the security of AI technologies and prevent their abuse. Even President Biden's administration recently instructed federal agencies to develop guidelines in this area.
