As AI-led attacks multiply, OpenAI launches a new cyber model
Anthropic says it will watermark text generated by its AI models
