Google previews Gemini agent tool-call semantic governance
Understand this story
Google has launched preview Semantic governance policies on the Gemini platform, deciding whether to allow an agent's proposed tool call before execution using two checks—natural-language policies and user intent—to block non-compliant calls from running. The policy currently supports only one enforcement point, Agent Gateway; enterprises need to configure agents so that all their outbound traffic passes through Agent Gateway; streaming model responses are not within the evaluation scope. Google says the feature is aimed at enterprises governing their agents and providing guardrails.
AI overview based on source reports · Updated 1 days ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Google Developers BlogFeaturedCheck what your AI agents do before they do it, with Semantic governance policies
Google Cloud's Semantic governance policies, in Preview in the Gemini platform, check each tool call an agent proposes before it runs against plain-language rules and against what the user originally asked for, enforced through a policy enforcement point such as Agent Gateway. A call must pass both checks; denied calls are replaced with a configurable denial message, and verdicts and the judge's rationale go to Cloud Logging. The post notes the LLM judge is probabilistic, recommends dry run mode before enforcement, and says streamed model responses are not evaluated.
Activity for this story
- Current comparable activity
- 3
- Peak comparable activity
- 5Nov 21 07:00
- Change in the past 24 hours
- -49%
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.