As the use of agent-based artificial intelligence becomes mainstream, Munich Re says the technology is expected to shape the scope, speed and precision of offensive and defensive cyber measures.
The global reinsurer’s recent report, Cyber Insurance: Risks and Trends 2026, highlights a world defined by geopolitical tensions, armed conflict and increased industry competition ahead, as well as an increasingly complex cyber threat landscape that requires decisive and proactive risk management.
The report states that agent AI will increasingly be able to plan and adapt to multi-stage operations, exploit vulnerabilities more effectively, learn from detection responses, and operate with minimal human input.
“Given that AI is already capable of generating deepfakes, authentic domain names and websites, and engaging in hyper-personalized social engineering and phishing, the existing attack surface will grow exponentially,” the Munich Re report explains.
The reinsurer continued: “As a result, an increase in synthetic content and characters, as well as misinformation, is expected to further undermine trust.
“Additionally, the AI models themselves will become targets of attacks and must be protected. Primary attack vectors will include on-the-fly injection and data poisoning, as well as the insertion of malicious data or instructions to manipulate output.”
The Munich Re report also noted that while agent AI can enhance parts of the cyber kill chain and lower the barrier to entry for attackers, the use of autonomous systems can also fundamentally change cybersecurity.
Risk owners’ views appear to reflect this sentiment. According to Munich Re, their overall views and expectations about AI are largely positive: only 23% of executives believe AI will have a negative impact on their business, 66% expect a positive impact, and 57% trust companies using AI.
On this topic, Munich Re concludes: “Despite all these technologies, the human element remains – both as a protective factor and as a potential gateway.
“Even with agent AI, humans won’t be completely replaced. So some of the current talk about agent AI seems more like hype.
“From Munich Re’s perspective, our experts expect that agency AI will have an impact on the frequency of attacks in the short term that exceeds their severity.
“Affected insurance types are particularly likely to include system failure and (C)BI, incident response, data recovery and cyber extortion – all of which are likely to be covered by first-party elements. In addition, the industry is likely to see more third-party losses from erroneous collection, breach of privacy, and media liability and technology E&O.”
The post-Agentic AI poised to shape offensive and defensive cyber measures: Munich Re appeared first on ReinsuranceNe.ws.