Is the Anthropic Usage Policy the Best Way to Ensure Safety?

Anthropic Usage Policy is now in effect, ensuring AI systems stop when connections drop. What does this mean for users?