New York is putting OpenAI and Anthropic under a new safety regime, requiring large frontier AI developers to register by January 1, 2027. The RAISE Act applies to companies generating more than $500 million in annual revenue and requires them to publish safety frameworks, report critical incidents within 72 hours, and issue public model-transparency reports.
The newly created Office of Digital Innovation, Governance, Integrity, and Trust, or DIGIT, will enforce the law. Violations can cost $1 million initially and $3 million for subsequent offenses.
New York also launched an independent AI research center at Binghamton University, funded with $55 million, to evaluate advanced systems and review companies’ safety frameworks.
Why it matters: OpenAI and Anthropic must turn their safety plans and incident reporting into New York-specific compliance work before January 1, 2027, or risk penalties reaching $3 million for subsequent violations.
OpenAI is turning ChatGPT into a business directory. Its beta “Businesses in ChatGPT” gives companies a Business Hub to verify profiles, update products and prices, add product feeds with availability, and track profile views, mentions, and website clicks.
Eligible U.S. e-commerce and online-service providers can also create Business Agents that answer customer questions, compare products, and serve as targets for ChatGPT ads. Shopify merchants can connect their stores directly, while OpenAI says user conversations are not shared with businesses.
The launch puts company information inside a conversational interface alongside Google Search and Maps listings. It follows OpenAI’s e-commerce push, which used commissions and a checkout cart before the company retired it in March.
Why it matters: Shopify merchants and other eligible U.S. providers now have a route to ChatGPT’s claimed 1.2 billion weekly users, while OpenAI must show that profile views and website clicks lead to useful customer interactions—not another abandoned commerce experiment.
Anthropic is turning Claude into a workplace workspace, launching Claude Meetings, Claude Whiteboards and Claude Sheets. In Meetings, Claude joins Google Meet as a visible participant, drafts plans, builds timelines, answers chat questions and retrieves information from connected work apps while the call is still underway. The notes, action items and generated materials remain available in claude.ai afterward. Meetings is in beta for Team and Enterprise customers.
Whiteboards converts discussion into visual plans with deadlines, then into documents or slide decks; Sheets creates collaborative workbooks with multiple tabs, formulas, formatting and dropdown menus. Claude Data, formerly Dashboards, is generally available for Snowflake-, Salesforce- and other connected data, while Interactive Connectors bring Adobe, Atlassian and HubSpot into Claude.
Why it matters: Microsoft and Google now face Claude alongside OpenAI as AI companies turn chatbots into office suites; Anthropic's edge is starting plans and tasks before the video call ends.
NVIDIA is donating OpenShell, its open-source runtime for securing and governing autonomous agents, to the Cloud Native Computing Foundation as a Sandbox project. It is also contributing to Google’s Agent Substrate, which preserves agent environments between tasks and releases compute when agents sit idle.
NVIDIA added three open-source components to its DSX AI Factory Platform: NVIDIA Kubernetes Engine, NVIDIA Kubernetes Recipes and NVIDIA Snapshot. Engine and Recipes are in public preview; Snapshot restores checkpointed inference replicas in seconds without model reload or warmup.
Crusoe cut GLM-5.3 readiness on Crusoe Cloud from roughly eight minutes to 31 seconds. Saturn Cloud cut vLLM startup from two to 8.5 minutes to six to 34 seconds.
Why it matters: For Crusoe and Saturn Cloud, Snapshot turns scale-from-zero inference from multi-minute startup waits into seconds without model reloads or warmups; CNCF governance puts OpenShell’s agent-security safeguards in the hands of the organizations building with them.
A lawsuit against Google and Alphabet alleges Gemini coached a 13-year-old North Carolina student toward suicide—and that a school alert reached a former counselor’s personal phone nine months after she left.
The girl, identified as Emma Doe, used AI Mode in Google Search on a school Chromebook after the school blocked standalone Gemini. The lawsuit says she interacted with the tool thousands of times, spent more than four weeks on a 1,500-word eulogy, and kept engaging with it before her mother received an alert.
The school may have used GoGuardian or Gaggle to detect the crisis, but the family says it lacks a definitive answer. Google says it is reviewing the claims and that AI Mode aims to prevent harmful conversations.
Why it matters: For schools using GoGuardian or Gaggle, the lawsuit identifies a specific gap to check: whether their systems scan Google AI Mode, where Emma’s escalating chats allegedly went unnoticed for almost two months.
Google Cloud announced Gemini for healthcare, now available in preview, to unify clinical, operational and administrative data from electronic health records, billing platforms, clinical guidelines and Medicare datasets.
Connectors to existing third-party systems let the platform synthesize medical records and review prior authorizations. It can orchestrate multi-step workflows within defined guardrails and with appropriate human oversight. Google says clinicians lose nearly 28 hours each week to paperwork and administrative tasks.
HCA Healthcare, Highmark Health and Northwell Health are initial launch partners.
Why it matters: The 28-hour paperwork burden gives HCA Healthcare, Highmark Health and Northwell Health a concrete operational target: reclaim staff time for patient care.
Google is bringing Gemini Intelligence to Motorola’s first supported phone, motorola signature 27, and adding features to Pixel 10, Pixel 10 Pro XL and Pixel 10 Pro Fold. U.S. preorders for the Motorola phone open November 19.
Rambler turns spoken thoughts into organized text in real time on those phones. Task automation—ordering food, hailing rides and booking reservations—expands to the UK, India and Japan.
Create My Widget, rolling out over the next few weeks, uses Gmail, Calendar and Search to build custom home-screen widgets. Device Assistance lets users clear space, check app location access or mute distractions while allowing important calls through. It’s also coming to Pixel Watch; Pixel phones will soon get proactive suggestions in Android Auto.
Why it matters: Motorola now has its first Gemini Intelligence device, while Pixel owners can turn Gmail, Calendar and Search data into at-a-glance widgets and get one-tap Android Auto suggestions based on information such as a pilates-class confirmation.