Google hat mit dem SynthID Detector ein neues Werkzeug zur Erkennung von KI-generierten Inhalten vorgestellt.
In Kürze
- Unsichtbare Wasserzeichen zur Identifikation von Medien
- Aktuell nur für Google-eigene Inhalte nutzbar
- Über 10 Milliarden Inhalte bereits markiert
Google’s New Tool: The SynthID Detector
Google has launched a new tool called the SynthID Detector. This tool is designed to help better identify AI-generated content such as images, texts, and audios. The key feature? Invisible watermarks! These are embedded into the media through small, invisible changes and are machine-readable. This way, it can later be determined whether, for example, an image of a camel in a snowy landscape was artificially created.
Limitations of the SynthID Detector
However, there is an important limitation: the SynthID Detector only recognizes content created with Google’s own tools like Gemini or Imagen. Content from other AI generators, such as Midjourney, remains undetected. This means that the detection of AI content is not as comprehensive as one might wish.
Current Accessibility and Future Prospects
Currently, the tool is only accessible to a select group of journalists, scientists, and media companies. A public version is not yet in sight. Google has already tagged over 10 billion pieces of content with this technology, which shows how serious the company is about this technology. The goal is to create more transparency in dealing with digital media and to combat disinformation.
Challenges and Future Development
Despite these advances, challenges remain. To be truly useful on a large scale, the tool needs to become more open and independent. The discussion about the detection of AI content remains exciting, and it remains to be seen how the technology will continue to develop.
Quellen
- Quelle: Google
- Der ursprüngliche Artikel wurde hier veröffentlicht
- Dieser Artikel wurde im Podcast KI-Briefing-Daily behandelt. Die Folge kannst du hier anhören.
