Marsh counts 105 client cyber incidents in late October, up from 48 in late May
Its cyber practice leader says AI helps attackers exploit familiar weaknesses faster. The incident counts do not isolate AI’s contribution.
Loading page…
Its cyber practice leader says AI helps attackers exploit familiar weaknesses faster. The incident counts do not isolate AI’s contribution.
Marsh Risk’s U.S. and Canada cyber lead Meredith Schnur says AI-enabled and agentic tools are helping attackers find and exploit familiar gaps—unpatched software, exposed systems, and weak identity and access controls—faster. She also points to rapid AI adoption and slower containment and response as drivers, making the increase a joint attacker-and-defender issue rather than evidence of a new vulnerability class. Marsh’s figures cover only its clients and do not identify how many incidents involved AI.
Marsh clients reported 48 incidents in the final two weeks of May and 105 in the final two weeks of October.
Schnur attributed a significant number of incidents to AI, but gave no estimate of how many.
The figures are not a measure of cyber incidents across all businesses.
Businesses reporting cyber incidents to insurance broker Marsh recorded more than twice as many cases in late October as in late May. In a Reuters interview published November 17, Marsh’s cyber practice leader linked the increase partly to AI-enabled attacks—and to organizations taking longer to contain attacks and respond.
Marsh’s clients reported 48 incidents in the final two weeks of May, compared with 105 in the final two weeks of October. Those figures describe incidents reported by the broker’s clients, not a count of cyberattacks across all businesses. They also do not isolate how many incidents involved AI.
Meredith Schnur, Marsh Risk’s U.S. and Canada cyber practice leader, told Reuters that attackers are using AI-enabled and agentic tools to identify vulnerabilities. She attributed a significant number of the incidents to AI, without putting a figure on that assessment.
The weaknesses she described are longstanding: software left unpatched, exposed systems and weak controls over identities and access. AI helps attackers find and exploit those openings at a much faster pace, according to her assessment. The concern is faster exploitation of existing security gaps, rather than a newly identified class of vulnerability.
Schnur’s explanation also puts the organizations under attack in the picture. She identified the speed and scale of AI-tool adoption, together with longer containment and response times, as the main factors behind the recent increase in incident volumes. Her account therefore links the rise to both attackers’ tools and defenders’ ability to react.
Schnur described AI as an amplifier of risk, rather than a separate peril. In that framing, the exposure remains rooted in familiar weaknesses; AI changes how quickly attackers can take advantage of them.
Loading discussion...
Join the conversation
Explain whether new tools or fixing familiar weaknesses should take priority.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.