📺 Google previews new Gemini features for TV at CES 2026 ↗
Google’s basically trying to make your TV feel less like a menu maze and more like - a chatty roommate. Gemini on Google TV is getting conversational search, show recaps, and “I forgot the title but I remember the plot” style discovery.
The surprisingly slick bit: you can ask for a topic deep-dive and get a big-screen, narrated, interactive explainer - then keep poking it with follow-ups. It can also rummage through Google Photos by “people or moments,” and apply artsy styles to photos/videos for slideshow-y memories.
And yes, the most practical feature might be the dumbest-sounding one: tell the TV “screen too dim” or “can’t hear dialogue,” and it adjusts the right settings without you digging through menus. It lands on select TCL TVs first, and needs Android TV OS 14+ plus an internet connection.
🗽 New York Enacts the RAISE Act Regulating Frontier AI Models ↗
New York just stepped into the “frontier model rulebook” club - as the second US state after California to pass broad rules aimed at big, powerful AI developers.
The RAISE Act leans hard into transparency and incident reporting: large AI developers have to publish info about safety protocols, and notify the state within 72 hours after certain safety incidents. There’s also a new oversight office inside the Department of Financial Services, which feels like an intriguing home for AI governance.
Enforcement teeth exist: the attorney general can bring civil actions for missing reports or false statements, with penalties scaling up for repeat issues. No private right of action, so it’s not open season for lawsuits from random passersby.
🎙️ Open AI’s upcoming AI gadget to be controlled by voice ↗
OpenAI’s gadget plans are sounding more “ambient assistant” than “another rectangle you stare at.” Reportedly it’s a device with cameras, microphones, and speakers that can observe what’s happening around you and answer questions - and it has no screen, which lands somewhere between elegant and mildly terrifying.
Because it’s screenless, the control layer is mostly voice. The reporting says OpenAI’s focusing on a voice model that feels more natural and can speak and understand at the same time - which reads as obvious until you remember how often voice systems stumble on basic turn-taking.
Timeline-wise, the voice model is expected sooner than the device itself, with the gadget targeted for a later-in-the-year launch at the earliest.
💸 AI-driven inflation is 2026's most overlooked risk, investors say ↗
Some investors are basically saying: everyone’s partying over AI-led markets, but the hangover could be inflation - and not the cute, “slightly higher prices” kind. The argument is that massive AI build-outs can push up costs via energy demand and pricey chips, especially as hyperscalers race to build data centers.
If inflation re-accelerates, central banks may stop easing - or even tighten - and that’s where the “uh oh” lives, because tighter money tends to hit high-valuation tech first. Rate pressure can also make funding huge AI projects more painful, which is a fun feedback loop if you’re not holding the bag.
The vibe is not “AI is bad,” it’s more “AI is expensive in the real economy,” and markets might be underpricing that.
🩺 OpenAI sees big opportunity in US health queries ↗
OpenAI published research claiming a huge volume of healthcare-related usage - more than 40 million people worldwide asking ChatGPT health questions daily, and healthcare prompts making up a noticeable slice of all messages.
In the US specifically, the framing is blunt: people use ChatGPT to help navigate a stressful healthcare system, and OpenAI’s not treating that as an awkward side effect - it’s treating it as a policy opportunity (which feels bold).
The report points to patterns like people trying to understand symptoms, lots of conversations happening outside normal clinic hours, and a chunk of messages focused on dealing with health insurance logistics. It’s very “this is already happening, so let’s shape the rules,” or so it seems.
🤖 At CES 2026, Everything Is AI. What Matters Is How You Use It ↗
CES is back in full “AI in absolutely everything” mode - but the more interesting takeaway is that AI as a label is losing its punch. When every gadget has a chatbot, the differentiator becomes software maturity and practical value, not the fact it says “AI” on the box.
WIRED’s take is basically: the market’s hit saturation, so execution wins. Smart glasses, wearables, smart home devices, health sensors - they can all do similar tricks, but the one that feels smooth (and doesn’t annoy you) is the one people keep.
There’s also this lingering suspense around OpenAI’s consumer-device strategy - everyone’s watching, nobody knows exactly what the “must-have” form factor is yet, and that uncertainty is kind of the point.
FAQ
What was the big theme at CES 2026 when “everything is AI”?
At CES 2026, the point wasn’t that AI is new - it’s that “AI” as a label is beginning to carry less weight. When every device claims to be “AI,” the winners are the ones that feel seamless and genuinely helpful in daily life. That’s why practical touches (like better TV discovery or fewer annoying menus) end up mattering more than flashy demos. Execution is the differentiator.
What new Gemini on Google TV features did Google preview at CES 2026?
Google previewed Gemini on Google TV upgrades like conversational search, show recaps, and discovery based on fuzzy plot memories. A standout feature is a big-screen, narrated, interactive “topic walkthrough” explainer you can keep refining with follow-ups. It can also search Google Photos by “people or moments,” apply artsy styles to photos or videos for slideshows, and handle settings tweaks through plain-language complaints.
How does Gemini on Google TV help when I can’t remember a show title?
Gemini on Google TV is designed for “I forgot the title but remember the plot” moments by letting you describe what you remember and guiding you toward likely matches. It also adds recaps, so you can quickly catch up without hunting through episodes or summaries. The overall goal is to replace menu-heavy browsing with a more conversational back-and-forth that keeps narrowing results as you clarify details.
Which TVs get Gemini on Google TV first, and what do I need for it to work?
Gemini on Google TV is landing first on select TCL TVs. The requirements mentioned include Android TV OS 14 or newer and an internet connection. In practice, that means older Android TV versions may not support the new Gemini experience, and offline use won’t cover the features that rely on cloud-based responses or content lookups. Availability is staged rather than universal on day one.
What is New York’s RAISE Act, and what does it require from frontier AI model developers?
New York’s RAISE Act is a broad law aimed at large, powerful AI developers, making it the second U.S. state after California to pass major “frontier model” rules. It emphasizes transparency around safety protocols and requires developers to notify the state within 72 hours after certain safety incidents. It also creates a new oversight office within the Department of Financial Services.
What do we know about OpenAI’s voice-controlled gadget, and why are investors worried about AI-driven inflation in 2026?
Reporting describes OpenAI’s device as a screenless, voice-controlled assistant with cameras, microphones, and speakers that can observe what’s happening and answer questions. Separately, some investors warn AI build-outs could drive “AI-driven inflation” via energy demand and expensive chips as data centers expand. If inflation rises, central banks could stop easing or tighten, pressuring high-valuation tech and making massive AI projects harder to finance.