Asia is racing to adopt AI agents. The rewards are clear – but so are the security risks

AI agents can be tasked to draft emails, write code and update documents, among other things. But the same autonomy that makes them valuable also creates vulnerabilities, warn cybersecurity experts.