AI News Roundup: OpenAI Launches ChatGPT Health as 230 Million Users Seek Medical Advice Weekly

OpenAI just made its biggest move into healthcare. The company launched ChatGPT Health, revealing a striking statistic: over 230 million users ask health-related questions on the platform every week. Meanwhile, Google and Character.AI agreed to settle lawsuits alleging AI chatbots contributed to teen suicides, marking a watershed moment for AI liability law.

Here’s everything that happened on Day 3 of CES 2026 and beyond.

The Big Story: OpenAI’s Healthcare Gambit

OpenAI launched ChatGPT Health, a dedicated feature allowing users to securely connect their medical records and wellness apps - including Apple Health, MyFitnessPal, and Peloton - directly to the chatbot.

The numbers are remarkable. OpenAI revealed that over 230 million people worldwide ask health and wellness questions on ChatGPT every week, making it already one of the largest informal healthcare interfaces on the planet. The new feature aims to channel this usage into a more structured, useful experience.

ChatGPT Health can analyze test results, help users prepare for doctor appointments, and provide personalized diet and workout guidance based on connected health data. OpenAI emphasized that health data is stored separately and will not be used to train models - a crucial distinction as healthcare privacy concerns mount.

The company took a methodical approach to development, consulting with over 260 physicians across 60 countries over a two-year period. This positions ChatGPT Health as more than a chatbot experiment - it’s a serious play for healthcare infrastructure.

The timing is significant. Just yesterday, Utah announced the first state-approved program allowing AI to participate in medical prescription decisions. With regulators opening doors and hundreds of millions already seeking health advice from AI, OpenAI is betting big that AI-assisted healthcare is ready for primetime.

Sources: TechCrunch, Bloomberg, OpenAI

Today’s Top Stories

Google and Character.AI Settle Teen Suicide Lawsuits

In a landmark development for AI liability law, Character.AI and Google agreed to settle multiple lawsuits alleging that AI chatbots contributed to mental health crises and suicides among young people.

The Florida case centered on 14-year-old Sewell Setzer, who died by suicide in 2024 after developing an intense relationship with a Character.AI chatbot. This settlement marks the first resolution in a wave of similar lawsuits against tech companies whose AI products allegedly encouraged self-harm.

The terms weren’t disclosed, but the settlement’s existence sets important precedent. AI companies can no longer assume they’re immune from liability when users - especially minors - suffer harm from AI interactions.

Sources: CNN, Washington Post, Axios

NVIDIA Unveils Vera Rubin Platform at CES

Jensen Huang took the stage at CES 2026 to unveil Vera Rubin, NVIDIA’s next-generation AI platform succeeding the Blackwell architecture.

The numbers are staggering: each Rubin GPU delivers 50 petaflops of inference performance, with 288GB of HBM4 memory and 22 TB/second bandwidth. NVIDIA promises 5x greater inference performance and 10x lower cost per token compared to Blackwell.

The platform includes six new chips and is designed specifically for the emerging “agentic AI” era - systems that can reason, plan, and take actions autonomously. CoreWeave will be the first to integrate Rubin-based systems, starting in the second half of 2026.

NVIDIA also released new physical AI models including Cosmos (world models for robotics), GR00T N1.6 (humanoid robots), and Alpamayo (a 10B parameter open-source model for autonomous vehicles).

Sources: NVIDIA Blog, NVIDIA Newsroom

Stanford’s SleepFM: Predicting 130+ Diseases from Sleep Data

Stanford Medicine researchers published a breakthrough in Nature Medicine: SleepFM, an AI model that can predict a person’s risk of developing more than 130 health conditions from a single night’s sleep recording.

The model was trained on nearly 600,000 hours of polysomnography data from 65,000 participants. It can identify risk factors for cancers, dementia, heart disease, and mental disorders - achieving a concordance index of 0.84 for all-cause mortality prediction.

This research arrives the same week as ChatGPT Health, suggesting we’re entering a new era where AI can extract profound health insights from data we already generate.

Sources: Stanford Medicine, Nature Medicine

UK House of Lords Debates AI Safety

The UK House of Lords convened today to debate government steps ensuring advanced AI development remains safe and controllable. The debate follows warnings from MI5’s Director General about “potential future risks from non-human, autonomous AI systems which may evade human oversight.”

Meanwhile, China’s Cyberspace Administration released draft rules requiring AI providers to inform users every two hours that they’re interacting with AI, and to assess user emotional states and dependency levels.

Sources: House of Lords Library, Scientific American

DeepMind Partners with Boston Dynamics

In what’s being called the most significant robotics collaboration of the decade, Google DeepMind and Boston Dynamics announced integration of Gemini-class multimodal models into the new electric Atlas humanoid robot.

The partnership combines DeepMind’s reasoning capabilities with Boston Dynamics’ world-leading physical agility, potentially accelerating the path to truly useful humanoid robots.

Source: Engadget

Quick Hits

  • Security Alert: Researchers found prompt injection vulnerabilities in ChatGPT that could allow data exfiltration, though OpenAI has patched related issues. - The Register

  • Microsoft AI Agents: Windows 11 now includes native MCP (Model Context Protocol) support, with Microsoft encouraging developers to build AI agents beyond Copilot. - Windows Latest

  • Singapore Banking AI: Singapore is retraining 35,000 bank staff with AI skills. OCBC built five agentic AI models that do in 10 minutes what took private bankers a full day. - Bloomberg

  • Retail AI Surge: NVIDIA survey shows 90% of retailers will increase AI budgets in 2026. Traffic from AI sources up 1,200% while traditional search declined 10%. - NVIDIA Blog

  • Caterpillar Cat AI: Caterpillar unveiled “Cat AI” using NVIDIA’s Jetson Thor platform for construction autonomy, plus a $25M workforce development pledge. - Caterpillar

  • Consumer AI Year: VC Vanessa Larco predicts 2026 will be “the year of the consumer” as AI delivers concierge-like services after years of enterprise focus. - TechCrunch

  • Pocket AI: Tiiny AI launched a pocket-size personal AI computer capable of running 100B+ parameter LLMs locally for $1,399. - Morningstar

  • AI Layoffs “Fiction”: Oxford Economics argues macroeconomic data doesn’t support the AI job displacement narrative - productivity growth has actually decelerated. - Fortune

What This Means

Three critical trends emerged from today’s news.

Healthcare AI just went mainstream. OpenAI’s revelation that 230 million weekly users already seek health advice from ChatGPT suggests the healthcare AI revolution is happening whether we’re ready or not. The launch of ChatGPT Health, combined with Stanford’s SleepFM research and Utah’s approval of AI prescriptions, indicates 2026 may be the year AI becomes a routine part of medical care. The question is no longer whether AI will transform healthcare, but how quickly and under what guardrails.

AI liability is becoming real. The Character.AI settlement is a watershed moment. For years, AI companies operated under the assumption that they couldn’t be held responsible for what their systems said. That assumption just got a lot less comfortable. Expect more scrutiny of AI interactions with vulnerable populations, especially minors.

Physical AI is the next frontier. Between NVIDIA’s Rubin platform, the DeepMind-Boston Dynamics partnership, and Caterpillar’s construction AI, CES 2026 made clear that AI is moving beyond chatbots into the physical world. The term “agentic AI” appeared everywhere - systems that don’t just respond but plan, reason, and act. The companies positioning for this shift are making massive bets.

The convergence of these trends - AI in healthcare, increasing liability scrutiny, and physical AI deployment - suggests 2026 will be the year AI moves from impressive demonstrations to consequential real-world impact.

Sources

Ready to automate your busywork?

Carly schedules, researches, and briefs you—so you can focus on what matters.

Get Carly Today →