Government of Canada Guide on the Use of Agentic Artificial Intelligence ' The main risks of agentic AI are not only about output quality, which could suffer fr
The Government of Canada has released a guide detailing the risks associated with agentic artificial intelligence. The guide highlights concerns beyond output quality, such as bias, errors, and harmful content. It also addresses risks related to unauthorized actions, unclear permissions, and accountability issues, noting that autonomous systems may take unintended steps if boundaries are not well-defined. AI
IMPACT Provides guidance for responsible deployment of autonomous AI systems by government entities.