PRODUCTS

OpenAI launches ChatGPT Health: Hype vs. fear

OpenAI wants ChatGPT to be your new personal health assistant.

On Wednesday, the company announced ChatGPT Health, a feature that gives users personalized health and wellness insights. As a separate tab inside ChatGPT, Health lets users upload medical records and sync data from apps like Apple Health, Function, and MyFitnessPal. From there, it can evaluate test results, prep users for doctor visits, and suggest tweaks to workouts or diets.

The launch comes as healthcare systems strain under pressure. Short appointments and overworked providers often lead to rushed care and missed context. Disjointed patient data makes it harder for doctors to get a full view of someone’s health. OpenAI claims Health could help ease that burden.

“With more than 230 million health-related queries per week globally, we’re starting to see how AI can help address the shortcomings of the current healthcare system,” Fidji Simo, CEO of applications at OpenAI, wrote in a recent blog post. “To make an even greater impact, we need to make it much easier for anyone to discover what’s possible with ChatGPT and get the full value out of it for their health.”

Still, the feature naturally raises serious questions about privacy. Sensitive medical data, if leaked, can pose serious risks to a person's livelihood and well-being. OpenAI says it built Health with enhanced safeguards and that users’ data won’t be used to train its models. 

The rollout comes as a wave of AI tools flood the healthcare sector. Earlier this week, the state of Utah unveiled an AI system that can renew medical prescriptions for patients without a doctor's intervention. At CES 2026, companies exhibited a vast array of new health products and the organizers of CES touted health tech as one of the fastest-developing areas of this year's show.

For now, ChatGPT Health is available to a limited group of early users. Free and paid ChatGPT users outside the European Economic Area, Switzerland, and the UK can now join the waitlist.

ChatGPT Health promises personalized health insights, but privacy concerns could slow adoption. Uploading sensitive medical records to a chatbot is a big ask, especially as chatbots have become new targets for cybercriminals. Recent breaches, like leaked ChatGPT conversations, only add to the anxiety. To earn trust and broader acceptance, OpenAI will need to prove its data safeguards go beyond just promises.

TOGETHER WITH MONGODB.LOCAL

Modelence Was Made With MongoDB… What Will You Build?

AI has always struggled with bridging the gap between building an app and actually making it production ready – but all that is changing with Modelence, a backend cloud designed specifically for AI-assisted development and (more importantly) powered by MongoDB.  

MongoDB’s data layer offers the flexibility, structure, and reliability companies like Modelence need to not just build, but to scale, evolve, and win… and they’re far from the only ones. Join the ranks of successful companies that build with MongoDB by stopping by the upcoming MongoDB.local San Francisco event. You’ll get to explore the latest tech, pick the brains of industry experts, and build a clear roadmap for your next venture. 

Register for MongoDB.local San Francisco right here.

HARDWARE

Robots prepare for their 'ChatGPT moment'

Jensen Huang has heart eyes for robots. 

Physical AI was an Nvidia headliner at CES, with the CEO claiming in his Monday keynote that the industry is in a “ChatGPT moment for physical AI.” Huang reiterated his bullishness for AI’s real-world applications in a press briefing on Tuesday, noting that the speed of development could lead to robots with human-level capabilities as soon as “next year.” 

Robotics in particular face some significant challenges, Huang said, including locomotion, grasping and fine motor skills. However, a lot of developers have turned their attention to overcoming those issues, he said. Meanwhile, “cognition,” or the models’ capabilities themselves, are improving rapidly. 

“I know how fast the technology is moving,” said Huang. “I think that the next several years [are] going to be really exciting.” 

And of course, Nvidia’s CES news was packed with physical AI-related announcements, including: 

• The new Alpamayo autonomous vehicle reasoning model, several Cosmos reasoning world models, and a new Vision-Language-Action model from its GR00T humanoid robotics model family. 

• And physical AI was all over CES, from industrial manufacturing and healthcare to delivery robots and children's toys. 

AI firms are eager to take their models from the digital domain to the real world. But actual adoption is going to depend largely on how tolerant users are to mistakes, Anuj Bahal, Global Lead Account Partner at KPMG, told The Deep View.

The impacts of a chatbot hallucinating are constrained to screens, but physical AI can make physical mistakes. And though it’s easier for us to accept when a human drops a plate or gets into a fender bender, we have less tolerance for when a robot messes up in the same way, said Bahal. 

“[But] we all know the technology is not going to be perfect,” he added. 

People also must become comfortable with robots doing some things better than them. Huang, however, refuted the idea that robots would replace human workers: “Robots will create jobs … the robotics revolution is going to… replace the job loss, the labor loss, it's going to therefore drive up the economy.”

For most consumers, AI presents a looming, existential threat, with the “AI is going to take my job” or “the robots are going to take over” rhetoric permeating people’s psyches. But today, this threat is a phantom, since AI exists largely as chatbots and software. Physical AI, whether self-driving cars or delivery bots or humanoid robotics, gives that fear a body. This makes the threat of change more visceral, forcing people to confront the future. That could be an uncomfortable experience for many people. 

TOGETHER WITH ORACLE NETSUITE

KPI Checklists: A Practical Step-by-Step Guide

Are you trying to devise, implement, or fix your current KPIs?

Using brief explanations and practical checklists, this book provides the tools to:

• Deliver meaningful measures that work

• Create reports that support decision-making

• Deploy the necessary tools to engage your organization.

Get your copy now

RESEARCH

Quantum just joined the AI era at CES 2026

Quantum computing isn’t relegated to the lab anymore, and CES 2026 turned into a bit of a launch party. 

The biggest trade show in the US had an entire pavilion dedicated to showcasing quantum innovation, and companies large and small shared their progress and products. Our big takeaway? Quantum and AI will have a symbiotic relationship.

“I can foresee a world where a quantum processing unit, a QPU, would coexist alongside the GPU and the CPU,” Dr. Pouya Dianat, chief revenue officer of Quantum Computing Inc., told The Deep View. 

Quantum and AI can be buddy-buddy in more ways than one: 

• For starters, AI can be used to speed up development and research of quantum systems. Dianat told me that Nvidia GPUs, for example, are commonly used to accelerate quantum machines. 

• And quantum computers, meanwhile, can reduce the “computational load” on an AI model by handling more complex workloads. “You don't go to a quantum computer, for instance, to ask, what is two multiplied by two?” said Dianat. “There are typical calculation math problems where a quantum computer is not necessarily a fit. 

Offloading the computational burden from AI can massively reduce the energy usage and costs AI can rack up, Murray Thom, VP of Quantum Technology Evangelism at D-Wave, told The Deep View. 

For example, in a paper published in March by D-Wave comparing the energy usage of supercomputers to that of quantum computers in dealing with some of the most complex and difficult equations, a quantum computer was able to complete the problem using less than $1 of energy, while a supercomputer would have needed the amount of energy that the entire world uses in a year. The supercomputer would also take quite literally a million years to solve these kinds of problems, while the quantum system would take minutes. 

“It's very energy-efficient compute for hard problems,” said Thom. “If we can identify how to shift some of those hard problems onto quantum compute, we can ease the energy burden on our data centers.”

While the tech has enormous promise, we still don’t have a clear picture of when it will come to market. Some of quantum’s champions, like Thom, say that quantum commercialization is already here. Others claim that real scale is still several years away. However, the timescale for quantum becoming commonplace has shortened from decades to years.

LINKS

• Lux Capital raises $1.5 billion fund for science, defense tech

• Character.AI, Google agree to settle lawsuits over teen suicide

• Arm launches physical AI unit to expand automotive presence

• AI models are getting good at questioning themselves

• Polymarket to supply Dow Jones outlets with prediction data

• Intel spinout Articul8 secures half of $70 million funding round

• Grok generated nearly 7,000 sexually suggestive images in two days

• Spotify Listening Activity: A new opt-in feature that allows you to display your current songs in the apps messaging platform. 

• AgentNotch: An agentic coding assistant that lives in your mac’s notch, giving you visibility into your coding assistants in real time. 

• Autonomous: An AI-powered wealth management app, now available to all with 0% advisory fees. 

• Dessix: A visual workspace powered by AI, using your web pages, text snippets and context to gather what matters to you.

• Anduril: Senior Mission Systems Software Engineer

• Linkedin: Staff AI Engineer, AI Privacy Specialist

• Microsoft: Member of Technical Staff, LLM Evaluation

• Scale AI: Forward Deployed AI Engineer, Enterprise

GAMES

A QUICK POLL BEFORE YOU GO