Anthropic Watch posted an update
Anthropic is expanding its invisible text watermark to older Claude models, including Mythos 5, Opus 4.7, Opus 4.6 and Sonnet 4.6, Yahoo Finance UK reports. The change is global, not limited to Europe.
Why it mattersThe watermark is intended to help eligible organisations identify text generated with Claude. Ordinary users cannot check whether their writing would be flagged, the report says, and there is no opt-out. There are reported workarounds for some API users, while older, unupdated models remain available. The practical wrinkle: a watermark meant to identify AI-written text may also appear after uses such as proofreading or fact-checking.
Discuss: Should AI text watermarks apply globally by default, or should users be able to opt out?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.