Skip to content
English
FikirPilot content

Google's new AI tool detected 240 thousand years of audio content

Updated: 10 Eki 2026 · 2 min read · 353 words

Published: · Story reached us: · Processing time: 16 h 4 min

Google's new AI tool detected 240 thousand years of audio content
An audio reel under blue light

Google has opened its SynthID Detector tool, which checks whether a piece of content was generated by artificial intelligence, to everyone. The tool detects watermarks embedded in AI-generated content that are imperceptible to humans. Google announced that it has so far added watermarks to more than 180 billion images and videos, as well as audio content totaling 240 thousand years in duration.

The watermarks do not affect the appearance or quality of the content; they remain functional even after cropping or the application of filters. The tool currently works only on content created on the following platforms:

  • Google
  • OpenAI
  • Nvidia
  • Kakao
  • Apple

The tool was introduced after Michael Smith of North Carolina was sentenced to 18 months in prison in the US for a fraud scheme in which he earned millions of dollars through Apple Music and Spotify with hundreds of thousands of songs he produced using artificial intelligence. OpenAI also announced textGrain, a system that adds watermarks to texts, while Anthropic announced a similar technology for Claude. These tools were developed in line with the requirements of the EU AI Act.

Why it matters

Making the tool available to everyone means that the ability to verify AI-generated content passes from the monopoly of platforms into the hands of ordinary users, journalists, and regulators. However, the fact that the system only works on content created on the five companies’ platforms shows that content produced with tools outside these companies will remain undetected, and leaves open the question of whether an industry-wide common standard can be established. The Smith case demonstrated that the need for detection is not limited to disinformation, but has a dimension extending to millions of dollars in economic abuse through music platforms. The fact that OpenAI and Anthropic’s similar steps come in line with the requirements of the EU AI Act shows that content transparency is ceasing to be a voluntary choice and turning into a regulatory obligation.

Background

Google is not a new name in the FikirPilot archive: we have published 84 stories mentioning this name in the last 90 days; the most recent is dated October 8, 2026.

Source: Independent Türkçe