Research & Development news

• ChatGPT Uninstalls Surge 295% After OpenAI Accepts Pentagon Contract

  04-03-2026 21:11 via eweek.com
  A Pentagon contract just cost OpenAI millions of its loyal customers in a single weekend.ChatGPT, OpenAI’s flagship product, saw a massive spike in uninstallations from US users, with a 295% day-over-day increase on Feb. 28. The surge followed public backlash after OpenAI signed a military partnership with the US Department of Defense, a deal that rival AI company Anthropic had previously declined over ethical concerns.How a Pentagon deal became a PR catastropheThe deal with OpenAI dates b
  
  Read more

• Meta AI Glasses Footage Reviewed by Humans, Including Intimate Moments

  04-03-2026 20:27 via eweek.com
  Meta’s AI smart glasses may be capturing more than users realize.Footage recorded by the devices — including highly intimate moments — can be reviewed by human workers helping train the company’s AI systems. According to the BBC, outsourced reviewers responsible for labeling the data say they have encountered deeply personal footage while doing the work.The news raises fresh questions about how recordings from the glasses are handled behind the scenes.A watchdog starts as
  
  Read more

• Google Home Gets ‘Live Search,’ Letting Gemini Analyze Your Cameras Live

  04-03-2026 20:00 via eweek.com
  Your smart home is finally starting to see the world the way you do. Google is rolling out a major update to its Home ecosystem, headlined by a feature called Live Search that lets Gemini AI analyze your camera feeds in real time. Instead of scrolling through grainy clips of things that already happened, you can now treat your cameras like a visual search engine.The standout addition, confirmed by Anish Kattukaran, chief product officer for Gemini for Home, allows users to ask about their p
  
  Read more

• Alibaba’s Qwen Tech Lead Steps Down After Major Model Push

  04-03-2026 19:56 via eweek.com
  Alibaba is dealing with the kind of AI headline no model team wants right after a release: a key technical leader behind Qwen says he’s stepping down.
  Junyang Lin posted a brief goodbye message on X and said he was leaving his role on the project without giving a reason. Alibaba hasn’t explained the change or named a clear successor publicly.
  According to TechCrunch, Lin’s post followed Alibaba’s release of new Qwen 3.5 “small models,” and Alibaba did not resp
  
  Read more

• Advertisement

• OpenAI, Google, and Alibaba Drop Faster, Cheaper Models

  04-03-2026 19:46 via eweek.com
  In the past 24 hours, three major players dropped new models, and none of them are trying to be “the smartest AI ever.” But they’ve all got a need for speed…Here’s what’s up:
  Google dropped Gemini 3.1 Flash-Lite.OpenAI answered with GPT-5.3 Instant.And Alibaba quietly shipped four Qwen 3.5 Small models that can run on your phone or laptop.
  These models are all optimized for the same thing: speed, cost, and running on smaller hardware.Here’s what each co
  
  Read more

• Apple Just Launched an Entire Lineup of AI-Powered Macs, iPhones, and iPads

  04-03-2026 19:25 via eweek.com
  Apple just dropped a small army of new products in a matter of days, and while the hardware is genuinely impressive, the real story running through all of it is artificial intelligence. From a $599 laptop for students to the most powerful MacBook Pro ever made, every single device Apple announced this week was built, in some meaningful way, around AI. That’s not a coincidence.The chip is the messageIf you want to understand what Apple is really doing with AI, start with the silicon.Th
  
  Read more

• Would ChatGPT Health Recognize Your Medical Emergency? New Study Raises Doubts

  04-03-2026 18:34 via eweek.com
  If you asked an AI chatbot whether your symptoms were an emergency, would it get it right? A new study finds OpenAI’s ChatGPT Health, a consumer chatbot designed to answer medical questions, misjudged more than half of serious medical emergencies.Published in Nature Medicine, the researchers found the system failed to recognize 51.6% of emergency cases, often advising patients to seek care within 24 to 48 hours instead of going to the emergency department.Stress-testing an AI health chatbo
  
  Read more

• Watch: This Chinese Humanoid Robot Gets Pushed, Kicked, and Still Keeps Dancing

  04-03-2026 18:20 via eweek.com
  Chinese robotics firm EngineAI has released new footage of its compact humanoid robot, PM01, performing a series of balance and recovery demonstrations designed to highlight control rather than raw speed.In the video, the robot dances before being deliberately pushed and kicked off balance. Instead of falling over completely, it absorbs the force, adjusts its posture in real time, and resumes its rhythm within seconds. The movements appear fluid and closely resemble human reactions to sudden dis
  
  Read more

• Advertisement

• GPT-5.3 Instant Arrives: Smarter Web Answers, Fewer AI Mistakes for ChatGPT

  04-03-2026 15:14 via eweek.com
  ChatGPT just got a conversational upgrade. OpenAI has rolled out GPT-5.3 Instant, a new version of its most widely used model designed to deliver smoother dialogue, more accurate responses, and stronger answers when pulling information from the web.The tech giant said the update focuses on improving the everyday experience of using ChatGPT, refining how the assistant responds so conversations feel more natural and consistently helpful.“This update focuses on the parts of the ChatGPT e
  
  Read more

• Google Launches Gemini 3.1 Flash-Lite, Its Fastest and Cheapest AI Model Yet

  04-03-2026 14:08 via eweek.com
  Google is pushing AI speed and scale further with a new lightweight model built for massive workloads. The company has launched Gemini 3.1 Flash-Lite, calling it its fastest and most cost-efficient model for high-volume AI tasks.Google says the model is intended for developers running high-frequency AI operations and real-time services that require fast responses across large volumes of requests.Built for scale, priced for productionGemini 3.1 Flash-Lite enters the Gemini 3 family as a streamlin
  
  Read more

• FinecoBank’s AI Bet: Faster Advice and Smoother Onboarding by 2029

  04-03-2026 12:14 via eweek.com
  Italy’s FinecoBank is betting that by 2029, AI won’t be a bolt-on feature. It’ll be part of how the bank finds customers, supports advisors, and moves money workflows faster, with fewer handoffs.
  According to FinecoBank, the board approved a 2026–2029 industrial plan that puts “AI-based technological innovation” at the center of growth across banking, investing, and brokerage. The plan reads less like a splashy product launch and more like a systems overhaul:
  
  Read more

Follow @R_and_D_News on Twitter!