new
🛡️ Smarter guardrails for safer AI interactions

We’ve added a new layer of protection to your AI-powered features to ensure safer, more responsible use.
What’s new
Messages are now automatically checked before they’re processed by the AI. These guardrails help block:
Harmful or sensitive content (e.g. violence, self-harm, or severe harassment)
Jailbreak attempts that try to manipulate the AI or extract internal instructions
How this helps you
You’ll enjoy more secure and reliable AI interactions—giving you peace of mind whether you’re using AI Assist or AI Agents.
No action needed. Just sit back and relax—these guardrails are already working behind the scenes for AI Assist and AI Agents.