Anthropic @AnthropicAI

We’ve trained Claude to flag potential misuse in conversations, a technique we call self-assessment. Right after it responds, Claude gives a preliminary safety assessment of the conversation. Since we began using self-assessment in June, it has surfaced misuse that our other safeguards weren’t designed to catch. Because Claude has already read the conversation, self-assessment is very efficient. And unlike classifiers, which target a discrete set of harms, it draws on Claude’s own judgment, so it gets more accurate as our models improve. Read more on our new Safeguards blog: https://t.co/18j6zdHub4

• User Says Anthropic Wrongly Tagged Them for Training Competitor Model2026-10-31
• OpenAI cuts funding target by $10B; valuation falls to $1.25T2026-10-31
• Anthropic blocks users touching training of larger models2026-10-31
• Anthropic watermark expands to more Claude models, with loophole2026-10-31
• Darkbloom idle-Mac AI network earnings fall to $1.23 a day2026-10-31