The Default Trap: Why Anthropic's Data Policy Change Matters
Read the terms of service. Don’t make assumptions. Don’t pick defaults. Yesterday, Anthropic quietly flipped a switch. If you're a Claude user, your conversations are now training data unless you actively say no. Not when you give feedback. Not when you explicitly consent. By default, from day one. Here's what changed: Previously, Claude didn't train on consumer chat data without your explicit thumbs up or down. Clean, simple, respectful. Now? Everything you type becomes model training fodder