Anthropic conducted training with Singapore's AI Safety Institute during the week of October 12, 2026 to instruct government and civil society experts on AI risk evaluation