Tool Calls

Tool calls are the mechanism that lets an AI model take actions outside its own text output, such as querying a database, sending an email, or running a transaction in a business system. Each call is a request the model makes to an external function, and the function's result is fed back to the model to shape its next step. For governance teams, this matters because it turns an AI system from something that only produces words into something that can change records, move data, or trigger real-world consequences, so permissions, logging, and human approval requirements need to be defined for each tool the model can reach.

1 item

ResearchGlobal2026-11-11

Visual Prompt Injection in Multimodal Agents Exposes a Tool-Call Control Gap

Researchers publishing in the ACM have demonstrated an adaptive visual prompt injection technique that manipulates AI agents processing untrusted images, websites, documents, or emails. The attack can cause agents to leak personal information or make unauthorized tool calls without requiring access to the model's internals. Enterprises deploying multimodal agents in document processing, email handling, or web browsing workflows face a control gap that existing text-based input filters do not address.

prompt injection multimodal agents agentic AI tool calls red teaming data privacy agent security