Microsoft and security vendors rush to address AI-era attack surfaces

A broad set of vendors and researchers focused this week on how AI changes security operations, from agentic defense to prompt injection and AI-powered attacks. Microsoft, Google, and others described new defense models that automate threat detection and remediation, while researchers warned that blind trust in agents creates dangerous new failure modes. Several reports highlighted prompt injection as the web-agent equivalent of XSS, and others showed AI-based phishing, voice impersonation, and malicious code generation becoming easier to scale. At the same time, new guardrail frameworks and defensive techniques such as 'context bombs' are emerging to slow down or trap malicious agents. This matters because AI is no longer just a productivity layer; it is becoming part of both the attack chain and the defensive stack. The central theme is that organizations must secure the agent, the data it can reach, and the actions it is allowed to take.

Sources

• The Real AI Threat Is Blind Trust — Dark Reading
• Google Bets 'Agentic Defense' Strategy Can Outpace Attackers — Dark Reading
• Prompt injection is becoming the XSS of the web agent era — Help Net Security
• The script, not the voice, is what makes AI voice phishing work — Help Net Security
• GPT-Red beat human red teamers on a prompt injection test — Help Net Security
• “Context bombs” can frustrate AI-driven attacks, researchers found — Help Net Security