With Siri AI and App Intents, we can surface content from apps. Can we add guardrails or configure what Siri can or cannot say when dictating content?
Guardrails for Siri + App Intents"
There are a bunch of guardrails, two prominent examples being confirmations, and authentication policy.
Intent confirmations enable gating risky intents behind explicit user confirmation. And intent authentication policy lets the developer require device unlocking before running an intent.
The system triggers these checks based on dynamic risk evaluation. But a developer can also configure these guardrails for certain situations.
For a deep dive, check out the WWDC26 session Secure your app: mitigate risks to agentic features.