According to Salt Security research, nearly half of enterprises reported that an AI agent took an unauthorized action over the past year. The finding highlights potential security and governance challenges as organizations increasingly deploy AI systems in their operations.