Here's your curated list of important insurance and AI news updates, critical use cases, and the latest innovations to help you stay ahead of the curve.
News: Are insurers being overwhelmed by AI-based data?
The root of it: In an interview with Insurance Thought Leadership, Dr. Michael Bewley, a geospatial data expert focused on insurance risk, says insurers face an explosion of AI-generated and third-party data that may complicate decision-making rather than improve it. While tools like supervised machine learning applied to aerial imagery can reliably assess property risk, insurers must carefully evaluate data quality, transparency, and uncertainty. The challenge is separating useful signals from noisy or unverified sources and integrating new AI capabilities thoughtfully into existing underwriting, risk, and catastrophe response processes.
News: A POTUS executive order guarantees shippers political risk insurance – and possible naval escorts
The root of it: A new executive order directs the US Development Finance Corporation to offer political risk insurance and financial guarantees for ships moving energy through the Strait of Hormuz after Iranian attacks prompted marine insurers to raise rates or cancel coverage. The administration also signaled the possibility of US Navy escorts for tankers. The move aims to stabilize energy markets and restore shipping confidence, as insurers and operators reassess risk in a region that accounts for about 20 percent of global seaborne oil trade.
News: Insurers are not responsible for Meta’s defense in social media addiction cases
The root of it: A Delaware court ruled that insurance carriers do not have to defend Meta in lawsuits alleging its social media platforms were designed to addict children and harm their mental health. The decision found that the claims involve alleged intentional conduct tied to how Meta (the parent company of Facebook and Instagram) designed and operated its products, rather than accidental harm typically covered by liability policies. The ruling shifts the burden of defense costs to Meta as thousands of similar addiction lawsuits move forward. This ruling applies only to the insurers' duty to defend, and not to indemnification.
News: Get ready for white-collar armageddon and the “new-collar” skills revolution
The root of it: A wave of viral essays and recent stock market reactions have sparked fresh fears that AI could rapidly displace white-collar roles, especially as Block CEO, Jack Dorsey (late of Twitter) announced a 40% workforce reduction tied to “intelligence tools.” Some analysts warn of a potential “human intelligence displacement spiral,” but others argue the threat is overstated. In insurance, AI is largely automating routine claims tasks while preserving human judgment, even as the AI boom fuels demand for new technical and “new-collar” infrastructure jobs.
News: US government civil rights unit rules that AI-generated job ads discriminated against workers
The root of it: The US Department of Justice settled with a Virginia IT firm, Elegant Enterprise-Wide Solutions, after AI-generated job advertisements unlawfully restricted applicants to certain visa holders. The DOJ said the ads limited consideration to workers with H-1B, OPT, or H-4 visas, violating the Immigration and Nationality Act by discriminating against U.S. workers. The company agreed to pay a $9,460 civil penalty, marking the eighth such settlement involving bias favoring visa holders.
News: Nvidia bets big on photonics to connect AI GPUs
The root of it: Nvidia is betting on photonics to improve how AI GPUs are networked within massive data centers. The company plans to invest $4 billion across optical technology suppliers Lumentum and Coherent, committing $2 billion to each. By replacing traditional electrical connections with light-based links, Nvidia aims to move data faster and reduce power use as AI systems scale to thousands of GPUs working together in increasingly large computing clusters.
News: What’s the best way to talk to a chatbot?
The root of it: Research suggests many popular prompt tricks – flattery, excessive politeness, or even threats – don’t reliably improve chatbot accuracy. Experts say better results come from clearly expressing the task and structuring prompts effectively. Useful techniques include asking for multiple answer options, providing examples, letting the AI ask clarifying questions, and avoiding leading language. Role-playing can help with brainstorming, but may reduce accuracy when a question has a single correct answer.
Read our 2025 State of AI Adoption in Insurance Report for insights and perspectives on AI adoption from insurance executives.