Anthropic introduced new safeguards that reduced unauthorized data harvesting from Claude by 90 to 99 percent. The defenses block foreign competitors from illicitly copying model reasoning to bypass safety rules.