Aug 19, 2026 · 46 min · 11 segments
William Collins and Eyvonne Sharp dig into the latest AI headlines, from the largest copyright settlement in American history to stolen AI models and invisible watermarks on Claude output. Plus, they…
Eyvonne SharpHostSo they're weaving... an imperceptible or imperceptible in whatever watermark directly into the actual text that Claude produces.
It's a signal embedded at the model level, essentially, like in the token choices themselves, if you will.
So that means it's gonna, like this watermark would essentially survive, you know, hey, I copied it out of here, put it in here, I pasted it here, and, you know, maybe some amount of editing, I suppose.
And basically, at the end of the day, anything Claude generates as a file is going to get digitally signed provenance metadata conforming to the...
So what is the C2PA standard? So in short, it's this industry way to cryptographically label the provenance of of digital media.
So people and systems can better determine authenticity and origin, I suppose.
I don't know if I, if that covers all of it, but the scope is like super broad.
And it applies whether you're hitting the Anthropic API directly or going through some third party cloud service.
And Another thing to add is the EU Act AI transparency thingy that rolled out worldwide is apparently the thing that stemmed this.
So they're weaving... an imperceptible or imperceptible in whatever watermark directly into the actual text that Claude produces.
It's a signal embedded at the model level, essentially, like in the token choices themselves, if you will.
So that means it's gonna, like this watermark would essentially survive, you know, hey, I copied it out of here, put it in here, I pasted it here, and, you know, maybe some amount of editing, I suppose.
And basically, at the end of the day, anything Claude generates as a file is going to get digitally signed provenance metadata conforming to the...
So what is the C2PA standard? So in short, it's this industry way to cryptographically label the provenance of of digital media.
So people and systems can better determine authenticity and origin, I suppose.
I don't know if I, if that covers all of it, but the scope is like super broad.
And it applies whether you're hitting the Anthropic API directly or going through some third party cloud service.
And Another thing to add is the EU Act AI transparency thingy that rolled out worldwide is apparently the thing that stemmed this.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.