Mayflower's Epstein: AI liability risks extend beyond cyber as adoption accelerates

By Isha Marathe Oct 13 (The Insurer) — AI-related liability may increasingly resemble traditional corporate claims involving negligence, directors and officers and employment practices rather than cyber breaches, said Mayflower Specialty CEO Jeremy Epstein. Speaking to The Insurer TV at ITC Vegas 2026, Epstein said the speed at which large companies have integrated AI into their core operations has surprised the specialist insurer since it emerged from stealth in July. A year and a half ago, much corporate activity around AI amounted to “a little bit of innovation theater,” Epstein said, with businesses testing and piloting tools. The technology has now become a material part of operations at major companies. Epstein said a cyberattack or rogue AI agent is not necessary to generate liability, with disputes potentially arising from ordinary business decisions, workplace practices and alleged negligence. That overlap with traditional directors and officers, errors and omissions and employment practices liability exposures has sharpened the need for explicit policy wording, he said. Some companies and brokers believe existing insurance provides adequate protection even without specifi