People Hate-Using AI? Insurance Agents' Agents; Midjourney Wants Your Body

Today's AI Outlook: ⛅️

AI Search Gets A Legal Reality Check

A German court preliminarily ruled that Google can be liable for false claims made by AI Overviews after two publishers said the feature labeled them as scams. The court treated the AI-generated summaries as new, independent statements, rather than neutral links to other people’s content.

Why It Matters

AI search has been leaning on the idea that disclaimers and “check the source” warnings can shift responsibility to users. This ruling cuts at that logic. When an AI system rewrites information into a confident answer, the court says the company may become the speaker.

The Deets

• The ruling barred Google from repeating the disputed claims.
• The court said AI Overviews go beyond traditional search results because they create substantive summaries.
• The case could become an early marker for how courts treat AI-generated search answers.
• The reporting noted that bad source links in Gemini overviews could turn small error rates into a large-scale liability problem.

Key Takeaway

Search engines used to just point. AI search speaks. That difference may come with a hefty bill.

🧩 Jargon Buster - AI Overviews: Google’s AI-generated summaries that appear above traditional search results and attempt to answer a query directly.

Americans Are Using AI While Side-Eyeing It

Pew found that 49% of U.S. adults now use AI chatbots, up from 33% in 2024, while 25% use them daily. The vibes, however, remain deeply unsettled: Only 16% expect AI to have a positive impact on society over the next 20 years, while about 40% expect a negative one.

Why It Matters

The industry loves adoption charts because they look like validation. Pew’s findings suggest something more awkward: AI is becoming unavoidable before it becomes trusted.

The Deets

• The top uses are search at 42% and work tasks at 38%.
• 63% of adults say AI is advancing too quickly.
• Majorities are worried about personal data.
• Skepticism remains strong even among younger adults.

Key Takeaway

People are using AI because it is everywhere, not because they have fully bought into the pitch deck.

🧩 Jargon Buster - Adoption: A measure of how many people use a technology, which does not necessarily mean they like it, trust it or understand it.

♟️ Power Plays

OpenAI Grabs A Transformer Heavyweight

OpenAI hired Noam Shazeer from Google, adding a Gemini co-lead and one of the co-authors of the landmark 2017 transformer paper, “Attention Is All You Need.” The move comes two years after Google reportedly spent $2.7B to bring Shazeer back from Character.AI.

Why It Matters

AI talent wars may be quieter than the loudest poaching sprees, but they are still shaping the frontier. Shazeer’s move is a signal that the biggest labs are still fighting over the people who know how to move models from impressive to dominant.

The Deets

• Shazeer joined Google in 2000.
• He co-authored the transformer paper that helped define modern chatbot architecture.
• He later founded Character.AI after Google declined an earlier chatbot-style idea.
• At Google, he served as a VP and Gemini co-lead.

Key Takeaway

In AI, talent is strategy with a badge swipe.

🧩 Jargon Buster - Transformer: The AI architecture behind most modern chatbots, designed to process language by weighing which words and context matter most.

State Farm Hands Agents The AI Keys

State Farm CEO Jon Farney told 19,000 sales agents their contracts are being replaced, with anyone staying past 2027 required to sign a new deal mandating daily AI use. The shift follows Progressive overtaking State Farm as America’s largest auto insurer.

Why It Matters

Insurance is becoming another proof point for AI’s corporate squeeze: companies want human teams to become more productive, while the tech itself reduces the need for some human touchpoints.

The Deets

• Progressive now sells more than half its policies direct, with AI-driven systems and no agent involved.
• State Farm agents who refuse the new arrangement can exit for $50,000 to $300,000.
• The company is trying to modernize a sales model that has long relied on local agents.

Key Takeaway

State Farm is trying to retrofit an agent-heavy business for an AI-first sales funnel. That is a culture change wearing a contract update.

🧩 Jargon Buster - Direct Sales: Selling policies straight to customers through digital channels, rather than through human agents.

🛠️ Tools & Products

Midjourney Trades Prompt Art For Body Scans

Midjourney revealed the Midjourney Scanner, a full-body imaging device that lowers users through water and a ring of ultrasound sensors to create body maps in about 60 seconds. The company plans to put the scanners inside its own spa locations starting in 2027, beginning with San Francisco’s Union Square.

Why It Matters

This is a strange, bold and very Silicon Valley expansion: an image-generation company moving into medical hardware with cold plunges nearby. If Midjourney can prove the scanner’s accuracy and utility, it could turn preventive health scanning into a consumer experience rather than a hospital errand.

The Deets

• The scanner uses underwater ultrasonic sensors.
• Founder David Holz said it can rival MRI-level detail in far less time.
• Midjourney built the machine with ultrasound-chip maker Butterfly Network.
• The first planned spa will include roughly 10 scanners, plus saunas, cold plunges and hot tubs.

Key Takeaway

Midjourney’s next canvas might be the human body, which is either thrilling or exactly how this sci-fi movie opens.

🧩 Jargon Buster - Ultrasound: Imaging technology that uses sound waves to look inside the body without radiation.

💸 Funding & Startups

World Models Are Getting Investor Heat

General Intuition is reportedly seeking $300M at a $2B valuation as investors chase startups building world-model technology. The company appeared in AI Secret’s daily roundup alongside other infrastructure and AI deployment moves.

Why It Matters

World models are becoming a major investor obsession because they aim to help AI systems understand environments, predict outcomes and plan actions. That matters for robotics, gaming, simulation and autonomous agents.

The Deets

• General Intuition is reportedly targeting a $2B valuation.
• Investor interest is rising around AI systems that can model physical or virtual environments.
• The funding pursuit sits inside a broader rush toward more capable agentic systems.

Key Takeaway

The money is following AI that can do more than answer. Investors want systems that can understand the room, even when the room is simulated.

🧩 Jargon Buster - World Model: An AI system that builds an internal understanding of how an environment works so it can predict what may happen next.

🧪 Research & Models

The Brain Gets A Better Decoder Ring

UC Davis published a Nature Medicine paper on BRAND, a machine learning platform that decodes neural signals into fluent language. The system translates attempted-speech brain activity into phonemes and words, and it works with off-the-shelf Blackrock arrays rather than custom hardware.

Why It Matters

Brain-computer interfaces have long faced a translation problem. Getting signals from the brain is one challenge, but turning those signals into reliable language is the breakthrough that makes the interface useful.

The Deets

• BRAND is being used across the BrainGate consortium.
• Trial participant Casey Harrell, who has ALS, has used it while working full-time as an environmental advocate.
• The platform is designed to be hardware-agnostic, meaning multiple implant systems could potentially use it.
• The key advance is the decoding layer that converts attempted speech into language.

Key Takeaway

The implant captures the signal. The decoder makes it speak.

🧩 Jargon Buster - Brain-Computer Interface: A system that lets the brain communicate directly with a computer, often by reading neural signals.

AI Doctors Move From Prediction To Preparation

In The Rundown AI, Rowan Cheung reflected on a conversation with Google DeepMind CEO Demis Hassabis, who said his confidence has “tightened” that AI could help cure all diseases within 10 to 20 years. Cheung’s practical takeaway: start collecting more personal health data now so future AI medical systems have a richer picture to work with.

Why It Matters

The AI health conversation is shifting from abstract promise to personal readiness. Wearables, workout logs, blood tests and biomarkers may become more valuable if medical AI systems can eventually analyze years of personal data.

The Deets

• Hassabis previously said AI could help cure all diseases within our lifetimes.
• Cheung argues that long-term health data could become unusually useful once medical superintelligence arrives.
• The suggested data includes fitness tracker history, workouts, sleep patterns and blood biomarkers.
• The point is preparation, not diagnosis today.

Key Takeaway

The future AI doctor may be smarter if your past self remembered to sync the wearable.

🧩 Jargon Buster - Biomarker: A measurable health signal, such as vitamin levels or blood markers, that can indicate what is happening in the body.

⚡ Quick Hits

• Anthropic said it is “very confident” its Mythos and Fable models will become available again in the coming days.
• Databricks launched new agentic tools at its Data + AI Summit, including LTAP for AI apps and analytics and CustomerLake, an AI-run customer data platform.
• Former White House AI adviser Dean W. Ball is joining OpenAI to lead Strategic Futures, a new frontier AI policy team.
• Amazon is selling its AI chips more directly as it tries to make Trainium a stronger Nvidia challenger.
• Snap spun off its AI video team into Dotmo after generative video costs became too heavy to keep inside the app.
• Google DeepMind warned that autonomous AI agents need controls similar to those used for potentially rogue employees with office access.
• The UK Cabinet Office is hiring an AI innovation leader to push an AI-first culture across the civil service.
• OpenAI researchers are working on ways to predict model failure rates before release, making AI deployment look more like stress testing.
• U.S. grid operators received federal backing for faster power-connection reviews as AI data centers increase electricity demand.

🧰 Tools Of The Day

• Google Gemini In Workspace Studio: Automates meeting prep by generating briefs, creating docs and blocking prep time before calls.
• OpenAI Codex: The agentic coding tool added Record & Replay, which helps create reusable coding skills.
• Perplexity Brain: A self-improving memory feature for Perplexity’s Computer agent.
• Adobe Firefly Studio: Adobe’s upgraded all-in-one platform for generating and editing creative work with AI.
• Crosby: An agentic law firm for sales teams designed to speed up time to signature.
• Mercury Command: An AI banking assistant that can answer cash-flow questions, follow up on invoices and turn natural language into account actions for approval.

Today’s Sources: The Internet, The Rundown AI, AI Secret