EXPLORINEER
en

Archive — US — techcrunch

Google’s new SynthID Detector can help spot AI slop

Google has introduced the SynthID Detector, a verification tool for identifying AI-generated content across various media formats, at Google I/O 2025. This comes amid a surge in AI-generated media, including a 550% rise in deepfakes from 2019 to 2024. While the tool is useful, it only detects content created with Google’s SynthID technology and has limitations, including the potential for circumvention.

Key points

  • Google launches SynthID Detector at Google I/O 2025.
  • The tool identifies AI-generated content in images, videos, audio files, and text.
  • SynthID Detector uses watermarking technology, with over 10 billion media pieces already watermarked since 2023.
  • The rise of AI-generated media is significant, with deepfake videos increasing by 550% from 2019 to 2024.
  • The tool is limited to Google’s SynthID specification and cannot detect content from other AI tools.
  • Google acknowledges that the SynthID Detector can be circumvented, especially with text.

Why it matters

  • The trend of AI-generated media has raised concerns about misinformation and digital trust.
  • In 2023-2025, many tech companies, including Microsoft, Meta, and OpenAI, are developing their own content watermarking technologies, highlighting a growing industry focus on combating AI-related deception.
  • The significant increase in deepfake technology and usage exemplifies the urgent need for tools like SynthID Detector.

The launch of the SynthID Detector is a timely response to the explosive growth of AI-generated media, addressing the pressing need for tools to ensure content authenticity, although it has certain limitations in detection capability.