A new dataset developed by an international team of researchers and published in Nature could revolutionize how human-recognition AI is developed and tested. FHIBE (Fair Human-Centric Image Benchmark) is the first large-scale, ethical, and diverse image dataset designed to combat bias and errors in computer vision systems.
Author: Mikolaj Laszkiewicz
Mikolaj Laszkiewicz
An experienced journalist and editor passionate about new technologies, computers, and scientific discoveries. He strives to bring a unique perspective to every topic. A law graduate.
Wandercraft has announced that its self-balancing exoskeleton, Atalante X, has received a second clearance from the U.S. Food and Drug Administration (FDA) in less than two years — expanding access to advanced, hands-free gait therapy for patients with spinal cord injuries (SCI) and multiple sclerosis (MS).
Fox News has published a report based on fake, AI-generated videos. The clips turned out to be entirely created by artificial intelligence. It’s yet another example of how easily misinformation can spread in the era of AI-generated videos without proper source verification.
Israeli startup Pulsenmore has received FDA approval to sell a home ultrasound device for pregnant women — marking the first time in history that a prenatal imaging technology has been authorized for use outside a clinical setting.
The biggest challenge in AI development isn’t a shortage of computing power, but rather infrastructure — power supply, cooling, and related systems. That’s according to Microsoft’s CEO, who says the issue affects every company working on artificial intelligence.
China’s campaign of economic espionage against the United States is shifting from cyberspace to inside American corporations. Instead of network breaches and data theft by hackers, the threat now increasingly comes “in white gloves” — through employees legally working at U.S. companies who transfer patents and technologies out of the country.
More and more American healthcare systems are implementing so-called “free” AI pilot programs. In theory, they are meant to improve efficiency — in practice, they…
The world’s biggest tech, e-commerce, and biotech companies are cutting jobs en masse, citing automation and the rise of artificial intelligence. But new research suggests that replacing people with algorithms too quickly could soon backfire.
As artificial intelligence features rapidly expand across web browsers like ChatGPT Atlas (by OpenAI) and Copilot Mode in Microsoft Edge, experts are raising serious concerns about user privacy and security — and research confirms those fears are not unfounded.
New research from the University of Cambridge, published in New Scientist, suggests that analog computers could train artificial intelligence systems up to 1,000 times faster than traditional digital machines — while consuming far less energy.