EU Urges Monitoring of High-Risk AI After OpenAI & Anthropic Incidents