🔴 So Anthropic just wrote a statement just to confirm all concerns about watermarking.
- It’s permanent and not removable.
- It will be applied to all new models from now on, and existing after December.
- They will apply it also to translations.
- They will provide a binary AI detection tool and API, that will determine whether AI was used or not, and not the extent or prior work from the user.
- It’s a code of conduct with the EU that they voluntarily signed.
- They decided to apply it globally instead of just inside the EU.
If you are using Claude for your university or a zero-AI policy job, stop using it right now. Either go back to manual writing or change to another provider without watermarking (like Grok).
Barışı destekleriz;
hukuksuzluğa destek vermeyiz.
Silahların bırakılmasını destekleriz; Demokratik siyasetin sınırlandırılmasını kabul etmeyiz.
Kardeşliği destekleriz;
hukukta çifte standardı kabul etmeyiz.