Claude by Anthropic, based in the USA, positions itself as an "ethical AI." Despite this claim, Anthropic has analyzed anonymized user conversations since the beginning. Furthermore, Anthropic changed its privacy policy in mid-August 2025 and will now use all user conversations by default to train its improved AI. This applies to past and new chats, regardless of whether the user has a paid subscription or not. Users can rate their conversations with "Like" or "Dislike," which ensures the chat will be included in the training of their AI models. As a US-based company, it is subject to US jurisdiction, which poses a clear risk to the data privacy of European users, as all data is stored on US servers.

Advantage: Claude is often praised for its ability to provide thoughtful and nuanced responses and is particularly strong in programming and creative writing tasks.

Privacy Verdict: The self-proclaimed "ethical" facade cannot mask the fundamental underlying privacy concerns. The use of data for training purposes for all users, regardless of subscription status, as well as its US location, represent significant drawbacks for individuals who value their privacy. Not recommended for business use.