Google's SynthID detector goes global and now flags OpenAI content too
The updated SynthID site identifies AI-generated media from Google, OpenAI and other providers.
Google has rolled out an improved SynthID content detector worldwide. The new site can identify AI-generated images, audio and text watermarked by Google, OpenAI and several other providers, not just Google’s own models.
Why it matters
- Platforms and publishers get a single place to check provenance.
- Watermark detection only works on content that was watermarked, so it is a signal, not proof.
- Related: OpenAI will watermark ChatGPT output by default, but only in the EU for now.
Source: Ars Technica