Google Limits AI Overviews After Health Safety Issues

Hey folks,

It’s Monday, so let’s quickly catch up on the some of the biggest and most interesting AI updates from the past few days.

Google Removes Some AI Summaries After Health Concerns

Google removed parts of its AI Overviews feature after a Guardian report found it was giving unsafe health advice. Some AI summaries suggested eating rocks for pica and drinking urine for cancer.

The issue was linked to AI summarising unverified web content. Google has now restricted AI-generated responses for certain health-related queries and said changes were made based on feedback.

OpenAI Launches ChatGPT Health in the US

OpenAI launched ChatGPT Health in the US, allowing users to connect electronic health records and wellness data. The feature supports data from providers like Epic and health apps.

OpenAI is testing the system with Northwestern Medicine. The company says connected health data will not be used to train AI models and remains protected under privacy rules.

NVIDIA Releases New Open AI Models at CES

NVIDIA introduced new open-source AI models, datasets, and developer tools at CES. This includes the Nemotron-4 models and resources for training custom AI systems.

The release also features blueprints for running AI efficiently on NVIDIA hardware. The tools are aimed at industries such as robotics, healthcare, and enterprise software.

RayNeo Introduces AR Glasses With Built-In eSIM

RayNeo unveiled the X3 Pro – Project eSIM prototype at CES 2026. The AR glasses include built-in eSIM and 4G connectivity, allowing them to work without a smartphone.

They support calls, music, real-time translation in 14 languages, and cloud syncing. The project is backed by funding from China Mobile and China Unicom.

AI Headphones Use Brain Signals to Track Focus

HyperX partnered with Neurable to showcase an AI-powered headset at CES 2026. The prototype uses EEG sensors in the earpads to monitor brain activity.

The system tracks focus, stress, and cognitive load in real time. It is designed for performance monitoring without implants or invasive technology.

Multi Model Comparison

With Geekflare Connect’s Multi-Model Comparison, you can send the same prompt to multiple AI models like GPT-5.2, Claude 4.5, and Gemini 3 at once. Their responses appear side-by-side in a single view, making it easy to compare quality, tone, and accuracy. This helps you quickly decide which model gives the best output for your specific task, without switching tabs or losing context.

Why AI Gets Expensive Without You Realising

Cheers,

Keval, Editor