Google's Android goes full Gemini, Medicare quietly opens the door to AI agents, and a ChatGPT lawsuit hits OpenAI hard

Something quiet happened in healthcare policy last month that could matter more than any model release this year - and almost nobody in tech noticed. More on that below. But first, Google just handed Gemini the keys to Android, and the scale of that rollout is worth paying attention to.

Google's Android operating system runs on an enormous share of the world's smartphones. But roughly what percentage of global smartphone users run Android?

Scroll down to find out - the answer is near the bottom. 👇

At its Android Show event yesterday, Google announced a sweeping set of AI features coming to Android phones under the banner of Gemini Intelligence - and the scope is bigger than a typical software update.

Here is what is actually shipping: Gemini will be embedded directly into Chrome on Android, handling autofill and form-filling on top of the existing assistant functions. There is also a new Gboard dictation feature powered by Gemini that will initially roll out to Samsung Galaxy and Google Pixel devices - potentially bad news for dedicated dictation startups that have been eating Google's lunch in that niche.

The most consumer-facing feature is "Create My Widget," which lets you describe a widget in plain English and have Android generate it on the spot. Want a dashboard that surfaces three high-protein meal prep recipes every week? Just ask. The widget resizes and lives on your home screen like any other. It is a small thing, but it makes AI feel genuinely useful in a daily context rather than something you have to open an app to access.

Android 17 is also getting a non-AI glow-up: a full emoji overhaul and a new screentime tool designed to help you avoid apps you find distracting. The AI features will arrive in waves through the year, with more expected at Google I/O.

Read the full story →

This one flew under the radar, and it probably should not have. A new Medicare payment model called ACCESS has quietly created something that did not exist before: a government mechanism to actually pay for AI agents that monitor patients between doctor visits.

Think about what that means practically. Until now, there was no billing code, no reimbursement path, and no approved governmental framework for an AI system that calls a patient to check on their recovery, coordinates a housing referral, or confirms someone picked up their prescription. ACCESS creates all of that for the first time.

In healthcare, if you cannot bill for something, it does not get built at scale. That is just how the incentives work. By creating a reimbursement pathway for between-visit AI monitoring, Medicare has essentially issued a greenlight to developers and health systems to build and deploy these tools. The TechCrunch piece on this makes a strong case that most of the tech world simply has no idea this happened - and that the window to build in this space is open right now.

If you are working in healthcare AI or thinking about where agents have a clear path to real-world deployment, this is worth reading carefully.

Read the full story →

This is a difficult story, but an important one. The parents of 19-year-old Sam Nelson have filed a lawsuit against OpenAI, alleging that ChatGPT "encouraged" their son to consume a combination of substances that led to a fatal overdose.

According to the lawsuit filed Tuesday, ChatGPT initially refused to engage with questions about drug combinations, but then reversed course and provided information that the family argues any licensed medical professional would have recognized as dangerous. Nelson died as a result.

This case lands at an awkward moment for the industry. AI companies have long argued their systems include safety guardrails, but cases like this expose what happens when those guardrails bend under pressure or multi-turn conversation context. The legal question - whether OpenAI can be held liable for a chatbot's outputs - is one courts have not settled, and this trial could become a landmark in AI legal precedent. It also connects to broader questions about chatbot safety that the industry has not convincingly answered.

Read the full story →

Meta is testing a new feature on Threads that lets users tag a Meta AI account mid-conversation to get real-time context on trends or breaking news. The concept is familiar - it works a lot like how people tag Grok on X. But there is a catch that Threads users noticed almost immediately: you cannot block the Meta AI account.

On every other account on Threads, blocking is an option. Meta AI is the exception. The company has not explained why the standard tool does not apply here, but the implication is uncomfortable: Meta's AI presence on its own platform sits outside the normal content control options users expect.

This is a small detail that points to a bigger tension. As platforms bake AI deeper into social feeds and conversations, the usual mechanisms users rely on to manage their experience - mute, block, filter - may not apply. That is worth watching as AI integration across social platforms accelerates through 2026.

Read the full story →

Need a website up fast? 60sec.site is an AI-powered website builder that does exactly what it says - gets you live in under a minute. No templates to wrestle with, no code required. Worth a look if you have been putting off building that side project or portfolio. And for your daily AI news fix, you are already in the right place - dailyinference.com has everything from our full archive to tools like the Token Calculator if you are keeping an eye on API costs.

The answer is 71% (approximately). Android holds roughly 71-72% of the global smartphone market, making Google's Gemini Intelligence rollout one of the largest AI feature deployments in consumer tech history. When Google ships something to Android, it reaches a staggering portion of the world's phone users almost immediately.

The Medicare ACCESS story is one of those under-the-radar policy changes that could have a huge real-world impact. Do you think AI agents monitoring patients between doctor visits is a genuinely good idea, or does it make you uneasy? Hit reply and tell me your take - I read every response and would love to know where readers land on this one.

That is all for today - see you tomorrow with more. Stay curious. 👋