Daily AI Briefing — May 13, 2026

Good morning. Here's your AI intelligence for Wednesday, May 13, 2026.

Clinical AI Crosses Into Daily Practice

Four major U.S. health systems have moved AI out of pilots and into clinical workflows, and the outcome data is beginning to arrive. The American Hospital Association profiles what scaled deployment looks like when 80% of U.S. physicians now use AI — double the 2023 rate. The pattern across health systems is consistent: AI embedded in triage, documentation, and diagnostic support — not replacing clinicians but restructuring how they spend time and reducing administrative load. What's changing now is the pressure to measure outcomes, not just adoption. Hospitals that deployed AI early are being asked by boards, payers, and regulators to show what actually improved.

The Intelligence Community vs. Deregulation

The administration's AI deregulatory posture hit internal friction this week. National security officials invoked the Defense Production Act to compel AI developers to share safety testing results with the government — a move that directly contradicts the White House's stated goal of stripping AI oversight requirements. The invocation signals a split inside the administration between agencies responsible for national security and political appointees pursuing a lighter regulatory touch. This is not a minor bureaucratic dispute. Whoever prevails determines whether AI safety disclosure is a voluntary commitment or a legal obligation for the country's most powerful model developers.

IMF: AI Enables Correlated Financial Attacks

The International Monetary Fund issued a formal warning that advanced AI dramatically reduces the cost and time required to exploit vulnerabilities in financial infrastructure. The specific concern: AI enables correlated attacks, where multiple institutions face simultaneous breach attempts optimized for the same shared systemic weaknesses. Existing bank stress tests were not designed to catch this pattern — they model firm-level risk, not coordinated AI-assisted exploitation across an entire sector. The IMF is pushing regulators in the U.S. and Europe to update financial stability frameworks before the next crisis, not after. Expect this framing to land at the Financial Stability Board before summer.

Autonomous Freight Goes Commercial

McLane Company — Berkshire Hathaway's $50 billion food distribution subsidiary — is running Aurora Innovation's self-driving trucks on live freight routes across the Sun Belt. This is not a pilot. It is commercial freight, real cargo, active routes. McLane's commitment represents one of the largest Fortune 500 bets yet on autonomous trucking as operational infrastructure. Aurora has been building toward this moment for years; what McLane provides is the scale validation the entire industry has been watching for. If this deployment holds over the next two quarters, it changes the conversation about when — not if — autonomous freight becomes the norm in long-haul distribution.

Factory Floor Convergence: Now Operational

45% of manufacturers now have live IIoT connectivity, and interest in large language models among manufacturers jumped from 16% to 35% in a single year. The convergence of AI, IoT, and robotics is no longer something manufacturers are planning — it is something they are running. The shift matters because these three technologies compound: real-time sensor data feeds AI models that direct robotic systems, closing a loop that human operators cannot replicate at scale. The next question is whether smaller manufacturers can access the same infrastructure, or whether the productivity gap between large and small industrial players widens into something structural.

Three States, Three Deadlines, No Federal Floor

Colorado's AI Act takes effect June 30. Connecticut's SB 5 is heading to the governor's desk. California is pursuing updated AI legislation — with overlapping, and in some cases incompatible, requirements across all three. Businesses operating across state lines face a compliance matrix with no federal preemption and no harmonization in sight. The practical pressure falls on any company that uses AI in consequential decision-making about employment, credit, healthcare, or housing — the highest-risk categories under all three frameworks. Legal teams are being asked to advise on three separate standards simultaneously, with enforcement timelines arriving before any of the open interpretive questions are resolved.

Real Estate: Industrial Wins, Residential Stalls

AI is not moving evenly through real estate. New JLL data shows 61% of institutional investors now use AI for market analysis — nearly triple the 2023 rate — but the gains are concentrated in industrial assets: warehouses, logistics hubs, and data center-adjacent properties. Residential markets are lagging, partly because the underlying data is messier and partly because the capital structures are different. The pattern is consistent with other sectors where AI adoption tracks data quality and deal frequency: early adoption yields go to the highest-volume, most-standardized transactions. The adoption gap in residential real estate is starting to look structural rather than temporary.

Google Expands Gemini Across the Smart Home

Google's Android 17 preview gives Gemini AI a broader smart home footprint: multi-step voice commands that can chain across devices, AI-powered camera intelligence that identifies context and responds accordingly, and Remy — a personal AI agent designed to learn household patterns and automate routines over time. The preview signals Google's intent to make Gemini the connective tissue across Android devices and compatible smart home hardware, not just a search or chat feature. Amazon and Apple are building toward the same position. The smart home platform war is now a Gemini vs. Alexa vs. Siri fight, with AI reasoning capability — not device count — as the differentiator.

Billion Into Precision Agriculture — But Not Evenly

More than $3.15 billion has flowed into AI-powered precision agriculture for crop micronutrients. John Deere covers 5 million acres with AI-assisted monitoring. Syngenta manages 70 million hectares. The tools are real and the yield data is compelling. But the capital is following existing scale: large commercial operations are capturing the technology gains while smallholder farmers — who grow a significant share of the world's food — are being left behind. As the farm bill debate continues in Washington, the equity question embedded in precision agriculture's growth is one that policy has not yet caught up with. The technology works; the distribution of its benefits does not match the distribution of global agricultural need.

What to watch today

• Connecticut's governor receives SB 5 — her signature or veto sets the tone for state-level AI safety legislation heading into the summer legislative window, with a dozen other states watching the outcome.
• Aurora's McLane routes are the most-watched autonomous trucking deployment of 2026. Watch for route expansion announcements and any public incident disclosures as commercial operations scale.
• IMF follow-up: whether the Financial Stability Board picks up the AI-cyberattack risk framing and begins translating it into updated stress-test guidance for member nations — a process that typically takes months but which the IMF is pushing to accelerate.