Information securityfromZDNET1 week agoHow indirect prompt injection attacks on AI work - and 6 ways to shut them downIndirect prompt injection attacks pose significant security risks to AI systems without requiring user interaction.