Gallagher Re flags AI aggregation risk

Gallagher Re published a white paper titled "Smart Systems, Blind Spots: Rethinking Insurance for the AI Era" in late March 2026, co‑led by Ed Pocock and Freddie Scarratt. (ajg.com) The paper explicitly lists AI failure modes that create new liability exposures: hallucinated outputs, poor automated decisions, discriminatory model behaviour, model drift and contaminated training data. (ajg.com) It notes courts and regulators are increasingly treating AI failures as the responsibility of the deployer rather than the vendor, while contractual vendor liability caps and limited indemnities commonly leave deployers financially exposed. (ajg.com) Gallagher Re maps which existing products (cyber, E&O, product liability, casualty) respond to particular AI loss‑scenarios and provides step‑by‑step guidance for designing AI‑specific covers that sit alongside those traditional lines. (ajg.com) The report identifies foundation models and widely adopted third‑party systems as vectors for correlated, cross‑sector losses that behave differently from geographically bounded natural catastrophe accumulations. (insurancebusinessmag.com) Target audiences named in the paper include underwriters and product leads, brokers and placement teams, reinsurers and portfolio managers, claims/wordings/legal teams, and risk managers and C‑suite technology leaders. (ajg.com) Gallagher Re has previously signalled that AI may eventually warrant a standalone insurance class—an evolution the firm raised publicly in mid‑2025 when discussing AI’s unique unpredictability and scale. (intelligentinsurer.com) Separately, Gallagher Re’s research referenced in the paper notes rising enterprise AI uptake—reporting that about 45% of businesses were using AI by the end of 2024—underscoring why the firm argues insurance structures must catch up. (reinsurancene.ws)