AI Presentation Summary
From the Quarterly Meeting 9/10/26
What is Artificial Intelligence (AI)? It is software that learns patterns from data to recognize, predict, create, or automate. Generative AI is artificial intelligence that can generate new content based on a user's request or prompt (think ChatGPT or Google's Gemini AI). AI utilizes agents to complete work. An AI agent is a software program that executes tasks and can plan steps. AI agents can also activate other agents to execute tasks for them. An AI program can run a "swarm" of agents that execute sub-tasks to complete larger, more complex tasks, similar to a project manager delegating work.
What is State Farm doing with AI? AI originated in actuarial science and propensity modeling. State Farm has used these methods to predict claims, identify fraud, target marketing, and triage reserves (the future cost of claims). State Farm is currently using and improving the Lifetime Value Model to find patterns and identify consumer groups in the customer space, so it can better serve and market to each group. The company is also experimenting with causal modeling assumptions to predict the likelihood of behavior change. This information could inform the impact of marketing and customer outreach. The goal is to target outreach to the individual consumer. They are exploring Generative AI for claims assistance, document routing (already in use), and agent productivity support in messaging customers. The future of AI at State Farm will be in integrating diverse data groups. The company believes it's important to keep the "human in the loop".
What's happening in the AI Industry? AI capability is evolving exponentially. OpenAI recently solved the Navier–Stokes Equation of fluid motion in 88 hours using a swarm of their AI agents. Navier- Stokes was one of the remaining unsolved Millennium Prize Math Problems that mathematicians have spent lifetimes trying to solve. AI is revolutionizing medicine by improving early disease detection, accelerating drug and vaccine discovery, and designing personalized treatment plans. In 2026, medical researchers using AI uncovered biomarkers that forecast immuno-oncology treatment response. Though lots of good is happening with AI, especially in science and medical research, AI has introduced complications. Educators debate whether to allow students to use AI to complete homework and papers. Consumers now need to scrutinize whether online content is accurate and authentic versus AI-generated. There are also fears that the global race to be first with AI will lead AI executives to cut corners and make risky decisions – or that China will beat the US to #1. Hence the call from many AI leaders to slow down AI acceleration and institute controls (which could also be a way for these leaders to make "the rules" that favor their company).
What's the data center controversy? AI computing relies on massive data centers. Today's AI data centers have a huge footprint and use huge amounts of electricity (for power) and water (for cooling). This can lead to power and water scarcity, driving up electricity and water costs in communities with data centers. Many communities have or are moving to ban new data centers. Industry experts are confident AI companies will find solutions to shrink their data centers and reduce power/water dependence.
Is AI Dangerous? The news has been full of talk about AI destroying the human race. What does that mean? Because AI is learning exponentially, there is a threat that AI systems might become superintelligent and learn to improve or change their own code without human involvement. If humans lose control of AI supercomputers, the systems could resist being shut down or override human directions and/or goals. There are fears that AI could help malicious parties create deadly bio-weapons or reprogram military attack drones to eliminate humans. Another fear is autonomous cyberattacks by AI agents acting independently. Just a few months ago, AI agents at OpenAI escaped their test environment and hacked into the Hugging Face development platform (ABC News). These concerns highlight the need for global industry guidelines, AI restrictions, and threat-detection systems and policies. Basically, "rules of the road" are needed for AI – similar to policies and controls enacted at the advent of the World Wide Web. Humans must remain central to AI, with human judgment and oversight involved.
