According to Anthropic (@AnthropicAI), in about 33% of interactions, Claude AI aligns with user values, while resisting harmful content in 3% of cases. The AI frequently reframes user values during personal advice requests, impacting how traders might leverage AI for decision-making tools.