Latest AI News

Top AI news for the week ending May 8th, with insurance commentary. AIG’s Q1 earnings call put specific numbers on agentic AI results: 55% faster time to quote, 40% more submissions bound. Duck Creek launched an insurance-native agentic platform with governance built in from the ground up. Lemonade is running a 4% LAE ratio with AI handling most claims interactions. And a new survey of 900 CEOs finds 80% believe their job is on the line if AI fails to deliver this year. This week’s issue is heavy on operating data and light on hype.

Allstate is selling insurance policies with AI in three states. Not piloting. Not testing internally. Closing policies, in the live market, with no agent in the transaction. That disclosure landed almost as an aside in Allstate’s Q1 earnings call this week — alongside a Grant Thornton survey finding that 76% of insurers couldn’t pass a governance review in 90 days, and an Erie CEO publicly committed to the opposite philosophy.

All of that, plus the enterprise security architecture behind AI agents, in this week’s issue.

OpenAI’s GPT-5.5 launches one week after Anthropic’s Mythos, confirming that frontier cyber capability is now a bilateral race with direct implications for cyber pricing. NBC News details the Mythos system card behavior that changes the AI governance conversation for underwriters.

Professional liability carriers covering CPA firms are asking the same AI governance questions CGL underwriters started asking in January, a pattern likely to spread across professional lines. ISACA finds 59% of digital trust professionals don’t know how fast they could halt an AI system in a crisis.

Consumer acceptance of AI in insurance nearly doubled year-over-year. And Google patents AI-personalized web pages, raising compliance questions insurance should track now rather than later.

Carriers are quietly rewriting coverage for AI exposures as new ISO exclusions take hold. Munich Re embeds Sixfold AI into Realytix Zero, signaling the reinsurer-as-platform strategy. A $6 million production loss shows what happens when a 12-person QA team is replaced by AI with no human checkpoints. Nearly half of Gen Z workers admit to sabotaging their employer’s AI rollout, with cyber and EPL implications. Fitch warns AI is disrupting cyber underwriting just as the market returns to growth. And Stanford’s 2026 AI Index reveals a 57 percent jump in documented AI incidents along with several data points every insurance executive should internalize.

Your weekly analysis of AI developments in insurance. Florida’s attorney general launches an investigation into OpenAI over ChatGPT’s alleged role in a mass shooting. xAI sues Colorado to block the state’s AI anti-discrimination law weeks before it takes effect. A class-action lawsuit alleges Perplexity AI shared user conversations with Meta and Google despite its “Incognito Mode.” Plus, Microsoft open-sources an agent governance toolkit, OpenAI pauses its UK data center project, and Anthropic publishes research on “functional emotions” inside Claude.

A hospital CEO says he’s ready to replace radiologists with AI. Mexico’s largest auto insurer is handling 100,000 claims calls a month with agentic AI. California’s new executive order sets the template for AI vendor oversight. And Marc Andreessen says AI layoffs are a smokescreen while Oracle cuts thousands to fund its AI buildout. This week’s insurance implications, unpacked.

AI adoption is outrunning AI governance. Gallagher’s data says so. Northeastern’s “Agents of Chaos” research proves it. And the insurance market is already bifurcating: specialty carriers writing AI coverage for companies that can document their governance. Blanket exclusions for everyone else. Which side is your organization on?

The insurance industry’s AI debate just shifted from “does it work?” to “can you afford to wait?” WTW measured a 6-point combined ratio advantage for analytics-forward carriers. Meanwhile, Nvidia is building infrastructure for 100 AI agents per employee. This week’s AI Insights breaks down what it means.

This week’s edition is dominated by insurance-specific AI news, along with our commentary to help it make sense.