Even if OpenAI and Anthropic boost their own safety measures, numerous models from other providers and open source vendors ...