SASSA plans to introduce voice recognition systems Ntuthuzelo Nene 30 April 2026 | 14:20Last September, the agency introduced a mandatory biometric system at all offices nationwide to combat fraud and verify the authenticity of recipients. SASSA national and Western Cape officials. Photo: Ntuthuzelo Nene The South African Social Security Agency (SASSA) said that it plans to expand its security features beyond fingerprint and facial recognition systems. Last September, the agency introduced a mandatory biometric system at all offices nationwide to combat fraud and verify the authenticity of recipients. The move followed significant challenges with SASSA payments, including widespread fraud, identity theft, and technical glitches. SASSA’s Executive Manager for Grant Administration, Brenton Van Vrede said the agency is now looking to incorporate voice recognition as part of its efforts to modernise its systems. "Because what we want to do is to create a trust centre where we capture all your biometrics. Noting there is no organisation that has every single biometric of every single person so that still has to be built. Home Affairs remains our foundational database for fingerprints and facials. So home affairs will ultimately, once everyone has replaced their green ID's and got smart ID's, the authoritative source on face and fingerprints." Get the whole picture 💡 Take a look at the topic timeline for all related articles.
Apr 30, 2026 · via ewn.co.za
Community worker Shaun Thompson and privacy campaigner Silkie Carlo, of Big Brother Watch, had been seeking to challenge the Metropolitan Police’s use of live facial recognition (LFR) under the European Convention on Human Rights, claiming the usage breached Articles 8, 10 and 11 of the ECHR. But their claim was rejected by the High Court, which ruled their human rights had not been breached by the technology’s usage, which was ““in accordance with the law”. The ruling comes after the Home Office recently ran a consultation on increasing use of facial recognition systems across the UK. Malcolm Dowden, a privacy expert with Pinsent Masons, said the decision would open the door to wider deployment of the tools by the authorities. “This case had been viewed as the first major challenge to deployment based on APP guidance on using facial recognition,” he said. “Its rejection is likely to fuel increased use of automated facial recognition, not only in policing but also - following the recent Home Office consultation - in areas such as border and immigration control.” Thompson had previously been misidentified by facial recognition systems in London, which had led to him being detained and questioned by police after it incorrectly identified him as his brother, who was on bail at the time. Lawyers argued that the use of the technology was a breach of privacy and pointed to a similar case involving South Wales Police in 2020 where the Court of Appeal had found the deployment of facial recognition systems had been unlawful, and was done so without adequate legal safeguards, leaving discretion to individual officers. Since that ruling in 2020 the College of Policing has issued authorised professional practice (APP) guidance on the deployment of live facial recognition technology which ensures that police forces address issues relating to
Apr 30, 2026 · via pinsentmasons.com
Abstract Video-based or image-based human activity recognition (HAR) via machine learning algorithms helps track, detect, and categorize users’ daily activities. It is usually formulated as specific research problems, such as fall detection, gait recognition, posture recognition, and gesture recognition. In practice, HAR is extended to a more generic group activity recognition (GAR) problem, focusing on multiple user (group) activities for various purposes, such as the recognition of social activities, autonomous driving, surveillance, and pedestrian crossing management. In this research, the formulation of the GAR problem aims to enhance the performance of the GAR model by tackling the research limitations in three aspects: (i) image quality may not always be guaranteed, resulting bias and inaccurate models; (ii) the working environments, in both indoor and outdoor, of the GAR model are dynamic, such as varying light conditions; and (iii) the number of people in the covered area may vary and the number of class labels for activities is large, leading to an extensive computing time to train a complex GAR model architecture. Therefore, a dynamic light image enhancement generative adversarial network has been proposed to enhance the image quality significantly. A multi-input image-enhanced generative adversarial network (MIIEGAN) has been proposed for generating high-quality additional training images, and a guided asymmetric depthwise separable convolution (GA-DSC) contributed to the model complexity and performance. These algorithms were evaluated in dynamic light environments to investigate their robustness and performance in varying light conditions that meet the nature of GAR in indoor and outdoor environments. Two benchmark datasets (The Volleyball Dataset and The Collective Dataset) were selected for the performance evaluation and comparison. Five backbones (AlexNet, VGG-16, Inception-v3, ResNet-50, and EfficientNetV2) were chosen as deep learning models for analysis. Nine existing methods were compared. Five variants of GANs and five variants of convolutional techniques were compared in
Apr 30, 2026 · via nature.com
DeepSeek "Open Eye" Sets the AI Community Ablaze: I Tested Its Capabilities with 12 Tricky Images to Find Its Limits Five days after DeepSeek delivered a powerful punch with V4, completely detonating the tech circle, Chen Xiaokang, a researcher in charge of multimodality within DeepSeek, posted the following on X and attached the text: Now, we see you. (Image source: Lei Technology) Yes, it means exactly what it says. While everyone was still amazed by the price and coding ability of V4, DeepSeek suddenly started testing the image recognition mode. The multimodal ability that the whole network had been discussing for a whole year has finally been implemented. The speed of this update really makes people wonder if Liang Wenfeng locked the development team in the computer room overnight to avoid being made into meme images of being irresponsible by netizens. It should be noted that this test is not a full - scale test but a small - scale gray - scale test. Only some users can see it in the official DeepSeek App or web version. At this time, in addition to the original quick mode and expert mode above the input field, there will also be a new image recognition mode button, marked with "Image understanding function is in internal testing". (Image source: Lei Technology) Unfortunately, none of my colleagues were selected for the gray - scale test. The number of people selected by the DeepSeek official was actually zero! Fortunately, I actually became the chosen one in ten thousand. Since it's such a coincidence, I'd feel a bit guilty if I didn't test it for everyone. This time, I carefully selected 12 pictures to let everyone see what DeepSeek can actually "see". Strong understanding ability, knowledge base needs to be updated Without further ado, let's start
Apr 30, 2026 · via eu.36kr.com
DietPal is an AI-powered nutrition tracking application designed to simplify calorie and fitness monitoring through conversational input and image recognition. Users can log meals, workouts, and body weight by interacting with a chat-based interface or by uploading food photos for automatic identification. The system processes this data to provide calorie estimates, nutritional breakdowns, and progress tracking over time. It also generates personalized diet recommendations based on user inputs and goals. Visual analytics, such as charts and graphs, are used to present trends in dietary habits and fitness progress. The platform is typically used by individuals aiming to manage weight, improve nutrition, or maintain structured health routines. DietPal reflects a broader trend in health technology that integrates artificial intelligence to reduce manual tracking effort and increase accessibility. Its primary function is to streamline dietary monitoring and support data-driven health and fitness decisions. AI Nutrition Tools DietPal AI Calorie Tracker Lets Users Log Meals And Track Nutrition Easily Trend Themes - Conversational Nutrition Interfaces — Natural language chat interfaces that allow logging and guidance through dialogue enable more intuitive and lower-friction user engagement with nutrition data. - Image-based Food Recognition — Photo-to-nutrient analysis that automatically identifies dishes and estimates portions opens possibilities for passive dietary tracking and richer food databases. - Personalized Diet Recommendations — AI-generated meal plans and nutrient targets tailored to individual goals and behaviors create opportunities for hyper-tailored nutrition strategies tied to outcomes. Industry Implications - Consumer Health Apps — Mobile wellness platforms that integrate AI tracking and analytics can shift user expectations toward continuous, personalized health feedback. - Fitness and Wellness Platforms — Integrated workout and nutrition ecosystems that combine conversational logging with progress visualization enable cohesive behavior-change offerings. - Food and Beverage Retail — Retailers and meal providers leveraging image recognition and personalized recommendations could transform product
Apr 30, 2026 · via trendhunter.com
“Suddenly the definition of what people meant when they were talking about AI changed underneath us,” Raiche said. “We went from talking about AI as statistical models for computer vision and image recognition, which … when we started our code of ethics and all that work, that was the world we were living in. Then suddenly we were in a world of generative AI.” Since then, Raiche and Vermont have established themselves as national trailblazers on AI. That code of ethics — which compares AI to power tools, instruments to be used by skilled operators to more effectively carry out tasks — has been adopted by many state and local governments across the country. Raiche has presented to other states setting up AI councils about his own state’s group, on which he sits. Raiche’s current title is chief data and AI officer, a role that recognizes the vital role data plays in enabling the state to work with AI. Often, that boils down to standardizing and streamlining procedures. “The prerequisite to actually doing the shiny AI work is to go through and update processes and to develop some consistency,” Raiche said. “It’s really business process engineering work that happens first, which will result in improved data quality, which unlocks the ability to leverage AI.” It’s that collaborative nature, with peers both in and outside the state, that Raiche said brings a positivity to his work. “We have folks who lean in to, ‘What could this look like?’ and who take the opportunity to learn,” he said. “And those people really encourage me when they trust that their leadership has their best interests and the best interests of Vermonters in mind, and then they lean in and learn something new and take a risk and come out better for it on
Apr 30, 2026 · via govtech.com
Facial recognition is now part of the security process at many U.S. airports, but travelers are not required to use it. At Transportation Security Administration (TSA) checkpoints, the technology usually appears at the identity verification podium before passengers reach the bag screening area. Instead of only handing over an ID for a TSA officer to inspect, a traveler may be asked to place a driver’s license, passport, or other acceptable identification into a Credential Authentication Technology device. At some checkpoints, that device can also take a live photo of the traveler. TSA says the system compares that live image with the photo on the traveler’s ID to help confirm that the person at the checkpoint matches the document presented. TSA describes this as facial comparison technology used for identity verification, and says the images are not used for law enforcement or surveillance. Still, many travelers may not realize they have a choice. According to TSA, passengers may opt out of facial recognition and request an alternative identity verification process. What The Facial Recognition Process Looks Like At checkpoints using TSA’s newer identity verification machines, the process usually starts when a traveler reaches the document checking station. The traveler presents an ID, and the machine checks the document’s information. If facial comparison is active and the traveler does not opt out, the device takes a live image at the podium. The system then compares that live photo with the image on the traveler’s government-issued ID. A TSA officer can review the result before allowing the traveler to continue to standard security screening. The Privacy and Civil Liberties Oversight Board’s 2025 report describes TSA’s checkpoint facial recognition system as a one-to-one matching system. That means it compares the live photo to the ID photo the traveler provided, instead of searching the person’s
Apr 29, 2026 · via travelnoire.com
In the last days of 2025, among tech commenters in rarefied corners of X, a new consensus emerged: AGI was here. By AGI, they meant artificial general intelligence—a term of art usually taken to denote computer programs that can match or exceed human capabilities at most economically valuable tasks. The immediate cause for the excitement was a series of updates to Anthropic’s AI coding tool, Claude Code, allowing it to complete complex programming tasks far more reliably, and with far less human supervision, than before. Taken merely as another productivity boost for software engineers, this would not be especially noteworthy. But understood as a stepwise improvement in a general system capable of performing almost any kind of computer-mediated work, it looked like a watershed: the automation not merely of coding as a specialized skill but potentially of any work performed with a computer. An Anthropic engineer posted that Claude Code was, itself, written almost entirely by a previous version of Claude Code. Its human overseers focused on “foundational architectural and product decisions,” while the AI implemented the solutions. Even the most in-the-weeds engineers on his team hardly wrote their own code anymore; they instead directed a team of agents. The long-foretold transformation of work seemed to have arrived. So, too, perhaps, had the man–computer symbiosis first described over 60 years ago by J. C. R. Licklider. “I have to constantly model the mind of [AI agents] living inside my laptop, and in doing so I become more like them. . . . I think in context windows. I become a cyborg, a hive mind of human and clauds,” wrote one anonymous tech poster on X. Finally, a reason to check your email. Sign up for our free newsletter today. And yet as excitement about the singularity grew in and around
Apr 29, 2026 · via city-journal.org
Abstract Background: Cataracts are an eye condition characterized by high prevalence and blindness-inducing potential, and effective approaches are required for their early diagnosis, underscoring the clinical significance of this study. Objective: This study aims to evaluate the performance of deep learning (DL) in cataract diagnosis and assess its potential as an effective tool for automated diagnosis, and compare the diagnostic accuracy of DL versus both machine learning and human experts. Methods: A systematic search was conducted in Web of Science, Embase, IEEE Xplore, PubMed, and Cochrane Library until April 1, 2025, for studies on image-based DL for cataract detection or clinical subtype classification. The included studies were assessed for the risk of bias (RoB) using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2). Bivariate mixed effects models were used for data analyses, and publication bias was assessed by Deeks’ funnel plots. Results: Sixty-three studies were finally included. The quality assessment indicated a high or unclear RoB in the patient selection (34 studies) and index test (44 studies) domains. Meanwhile, in the reference standard domain, the risk of bias was high or unclear in only 2 studies. Image-based DL achieved a sensitivity of 96% (95% CI 0.95‐0.97) and a specificity of 98% (0.96‐0.98) for cataract detection, with an area under the ROC curve (AUC) of 0.99 (0.98‐1.00). For cataract classification, the sensitivity and specificity of image-based DL were 94% (0.93‐0.96) and 97% (0.96‐0.98), respectively, with an AUC of 0.99 (0.98‐0.99). Despite the strong overall performance, the model’s generalization capability was challenged by its lower performance observed on independent external datasets (detection: sensitivity 87%, specificity 93%; classification: sensitivity 89%, specificity 90%), potentially attributable to domain shift between the training and validation data. Conclusions: Image-based DL has demonstrated high precision in the detection and classification of cataracts, showing potential advantages over traditional machine learning
Apr 29, 2026 · via jmir.org
New Orleans Police Continuing to Use Live Face Recognition Despite City Law Subscribe to the Free Future Newsletter Free Future home A troubling New Orleans face recognition program revealed by a Washington Post investigation last year continues to operate despite violating a city law and a claim that it has been paused, and is being used in more problematic ways than first reported, according to emails obtained through an activist’s public records requests. If allowed to stand, this one-of-a-kind program would represent an alarming new expansion of how face recognition is used in America. Face recognition is used by a number of police agencies around the nation, but almost always to discover the identity of an alleged perpetrator who appears in a photo. New Orleans police officers, however, have used — and appear to be continuing to use —live face recognition through an organization called Project NOLA. That means that they can do: - Real time monitoring. The systems operators can set the system to issue instant alerts when someone on a watch list appears on some of the system’s 5,000 camera system (the newer ones, which have face recognition functionality embedded within them). - Search. The operators can also enter a name or image and pull up the video of every time the person appears in any of the recordings the system has made (Project NOLA says it retains them for 30 days). - Tracking. The retroactive construction of a person’s movements across the city over time, which can be highly revealing of how they live their life. - Relationship identification. Such a system also enables even more invasive social network analysis by tracking a person’s repeated co-appearances in a video with another person. There has been no reporting that Project NOLA is doing this in automated fashion, but
Apr 29, 2026 · via aclu.org
Announcing a new publication from Opto-Electronic Advances; DOI 10.29026/oea.2026.250263 SHANNON, CLARE, IRELAND, April 29, 2026 /EINPresswire.com/ — Announcing a new publication from Opto-Electronic Advances; DOI 10.29026/oea.2026.250263 For centuries, the manipulation of light has relied on bulky, heavy components like glass lenses and prisms. From the camera in a smartphone to the microscopes in a laboratory, traditional optics are constrained by physical laws that dictate the size and thickness. However, a revolution is underway with the advent of “metasurfaces”. They are ultra-thin, flat optical devices composed of millions of sub-wavelength structures that can bend, focus, and filter light at will, in ways that natural materials cannot. This has the potential to shrink bulky cameras down to the thickness of a sheet of paper. Yet, designing these metasurfaces is a task of immense complexity. A single device may contain millions of nano-pillars, each requiring precise engineering and fabrication. Navigating this infinite design space using traditional human intuition or standard computer simulations is a significant bottleneck. This review article explores how artificial intelligence (AI) is shattering that bottleneck. Just as AI has transformed language processing and image recognition, it is now reshaping the field of optics, giving rise to “AI-Assisted Metaphotonics.” Deep learning algorithms are proving to be the perfect partner for nanophotonics. Where a human engineer might take weeks to simulate a single design, AI-powered “surrogate models” can predict how light will interact with a nanostructure in milliseconds. More impressively, AI enables “inverse design.” Instead of an engineer guessing a structure and checking if it works, they can simply tell the AI the desired optical property, such as a specific color or focal length, and the AI generates the exact complex geometry required to achieve it. But the synergy goes beyond just design. The review highlights how AI is being integrated into
Apr 29, 2026 · via desmoinesregister.com
In summary: - PCWorld tested Windows Hello facial recognition security by attempting to fool the system with photos, finding it effectively resisted all spoofing attempts including high-resolution iPad images and printed photos with eye holes. - Windows Hello uses IR cameras and 3D depth mapping technology to distinguish real faces from flat images, storing facial landmark data locally rather than actual photos for enhanced privacy and security. - The system proves significantly more secure than older facial recognition methods and traditional PINs or passwords, requiring sophisticated 3D facial replicas to potentially bypass its advanced biometric protection. Long ago, I had an Android phone with an early facial recognition sign-in feature… and someone could unlock my phone just by holding up a photo of me. Yeah, it was bad. Fast forward to 2025 and we have Windows Hello facial recognition sign-ins for PCs. Microsoft talks a big game about how secure it is, that Windows Hello can’t be easily tricked, that it’s better than a traditional PIN or password, and that it’s as secure as Apple’s Face ID. But is it really? I ran an experiment and tried to fool it. Here’s what happened when I put facial recognition to the test on my PC. How I tried to fool Windows Hello If someone wanted to fool facial recognition biometrics, they’d probably do it using a photo of your face. So that’s just what I did—I took a photo of myself (available online), put it on an iPad, and held it up in front of my face. My Windows Hello webcam wasn’t fooled for a second. In fact, Windows Hello doesn’t even see flat pictures as faces! While the Camera app on Windows does register it as a face, Windows Hello knows better. Despite holding up a high-resolution image of my
Apr 29, 2026 · via pcworld.com
Disney has rolled out facial recognition software at its theme parks in a fraud crackdown that has led to warnings from privacy experts. Pictures taken at certain entrances to Disneyland and Disney California Adventure in Anaheim are compared to images taken when a guest first used a ticket or annual pass, the company said, arguing that the technology can be used to stop fraud or potentially to prevent the misuse of annual passes. Facial recognition is optional and not in use at all entrances, Disney said. The data is deleted after 30 days except in cases where it “must be maintained for legal or fraud-prevention purposes”, according to the company’s website. Yet some experts have raised privacy concerns as the use of facial recognition software spreads throughout society. “The normalisation of facial surveillance is really problematic,” Ari Waldman, a professor of law at UC Irvine, told the Los Angeles Times. “We can’t go around life hiding our faces, so this isn’t just next step in surveillance; it’s qualitatively different. In a world of facial recognition, when people leave their house, it automatically means they’re identified.” Major sporting and entertainment venues have been using facial recognition software for what they say are security reasons. In Major League Baseball for example, many franchises allow fans to partake in “Go Ahead Entry”, which enables them to enter without a ticket as long as their faces are scanned. Madison Square Garden, the New York venue owned by the billionaire James Dolan, has been criticised for allegedly using the technology to keep track of critics of the businessman. The American Civil Liberties Union is among the groups that have previously raised the alarm about how prevalent the software is becoming in public life. Most fans at Disneyland recently paid little attention to the signs alerting
Apr 29, 2026 · via thetimes.com
Each edition of Research Bytes showcases the scholarly and professional achievements of Drexel School of Computer and Information Sciences (SCIS) faculty and students. Research Accomplishments Recent Grant Awards Computer Science Assistant Professor Feng Liu, PhD received an in-kind gift of $32,000 in GPU hours from the NVIDIA Academic Grant Program to support autonomous driving AI research. Learn more. Liu is also a Co-PI on two recently awarded 2026 Longsview Fellowships. The first, awarded to his colleague, Drexel School of Engineering professor Fernanda Campos da Cruz Rios, PhD (PI), will support their research on automated detection of façade-level material to advance urban material stock and flow analysis. The second, awarded to the PI, Drexel professor Matthew McDonald, PhD (chemical and biological engineering), will be focused on using generative AI for organic crystal structure prediction. Computer Science Assistant Professor Li “Harry” Zhang, PhD is a Co-PI on a recent 2026 Grimes Family Faculty Award with School of Engineering colleague Zhiwei Chen, PhD (PI). The award will support their research on neurosymbolic learning for safe autonomous driving under ambiguous trust. Computer Science Associate Professor Edward Kim, PhD received a grant of $11,025 from trading firm Hard Eight Trading, LLC that will support PhD research on comparative analysis of machine learning models for gold futures forecasting. Honors & Recognition Professor and Information Science Department Head Helena Mentis, PhD received the Association of Computing Machinery Special Interest Group on Computer-Human Interaction (ACM SIGCHI) Lifetime Service Award. This award highlights Mentis’s remarkable contributions to the growth and success of SIGCHI over many years. Learn more. First-year information science PhD student Tzu-Yu Weng and Information Science Assistant Professor Karthik S. Bhat, PhD received an Honorable Mention Award for their paper “A Blessing and a Challenge: Unpacking Boundary Ambiguities Experienced by Caregivers of Older Adults” that will be
Apr 29, 2026 · via drexel.edu
Image: Cottonbro studio - Pexels A woman strolls into a grocery store, thinking about grabbing some apples. Before she even reaches the produce aisle, a security camera has scanned her face. Whether the system is checking for shoplifters or simply logging her arrival, her face has joined a digital ledger, a trace she can’t easily erase. Retailers, banks, airports, stadiums and office buildings are doing the same. But what if the woman’s facial information is stolen or misused? If a cybercriminal steals her password, she can change it. If they acquire her credit card number, she can cancel the card. But she can’t reset or revoke the appearance of her cheekbones. Facial recognition systems don’t keep actual images. They convert a face into a mathematical template that maps the positions and proportions of the face’s features. When another camera scans a person later, the system checks their live face against these templates to confirm an identity. In my work as a cybersecurity professor at Rochester Institute of Technology, I have found that even though templates are more secure than photos – which anyone online can capture and manipulate – templates, too, can be stolen. Once that happens, these digital keys create a lifelong vulnerability. If a facial recognition database is breached, the “locks” that a template opens – accessing a bank app, getting through security at an airport, entering an office building – can’t be reset. A person’s face is permanent, and so is the threat. The threat isn’t theoretical. Biometric data has been stolen in data breaches. In 2024, biometric data from a facial recognition system used at bars and clubs in Australia was hacked. And in 2019, biometric data from a pilot facial recognition system set up by U.S. Customs and Border Protection was breached in an attack
Apr 29, 2026 · via digitalinformationworld.com
© 2026 Barchart.com, Inc. All market data is hosted and powered by Barchart.
Information presented is provided 'as-is' and solely for informational purposes, not for trading purposes or advice, and is delayed. No representations are made by Barchart as to its informational accuracy or completeness.
Apr 29, 2026 · via scottcoop.com
Abstract The United Nations’ Sustainable Development Goals, SDG 12: Responsible Consumption and Production, and SDG 13: Climate Action highlight the importance of environmental conservation and reducing pesticide use. Early and accurate pest identification is essential for implementing targeted pest control measures, which helps reduce unnecessary and incorrect pesticide use. While effective pest recognition and classification are crucial for ecological research and biodiversity conservation, traditional methods remain labor-intensive, time-consuming, and dependent on experts. Several deep learning techniques have been introduced in recent years, leading to more efficient and accurate identification and classification of crop pests. This research presents a structurally adapted DenseNet model for multi-class pest image classification based on dense connections. The model is fine-tuned through hyperparameters involving dense blocks and transition layers to perform consistently across three different datasets, including the IP102 dataset, which contains over 75,000 images of 102 pest species. The study also addresses dataset imbalance to prevent biased outcomes by deep learning models. The proposed structurally adapted model for fine-grained classification achieves 82.69% accuracy and 81.45% F1 score on the IP102 dataset, complementing existing advanced methods. Similar content being viewed by others Funding Open access funding provided by Manipal University Jaipur. The authors received no funding for this work. Author information Authors and Affiliations Corresponding author Ethics declarations Competing interests The authors declare no competing interests. Additional information Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Rights and permissions Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
Apr 29, 2026 · via nature.com
SenseTime, a Chinese AI company best known for its facial recognition technology, released a new open source model on Tuesday that it claims can both generate and interpret images far faster than top models developed by US competitors. SenseNova U1 could help the company reclaim lost ground after it slipped from its place among the leading players in China’s AI development race. The model’s secret sauce is its ability to “read” images without translating them to text first, speeding up the process and reducing the amount of computing power required. “The model’s entire reasoning process is no longer limited to text. It can reason with images as well,” Dahua Lin, cofounder and chief scientist at SenseTime, said in an interview with WIRED. Lin, who is also a professor of information engineering at the Chinese University of Hong Kong, says that models capable of processing images directly will enable robots to better understand the physical world in the future. Like DeepSeek's latest flagship model, SenseTime says U1 can be powered by Chinese-made chips. “Several Chinese domestic chipmakers have finished optimizing compatibility with our new model,” Lin says. On release day, 10 Chinese chip designers, including Cambricon and Biren Technology, announced their hardware supports U1. That flexibility matters because US export controls restrict Chinese firms from accessing the world's most advanced AI chips, particularly those used for training, which at this point are primarily developed by Western companies like Nvidia. “We will continue to push for training on more different chips,” Lin says. But he also acknowledges that SenseTime “may still need to use the best chips to ensure the speed of our iteration.” SenseTime released U1 for free on Hugging Face and GitHub, another sign of how Chinese companies are becoming some of the most active contributors to open source AI.
Apr 29, 2026 · via wired.com
If you want to visit the “Happiest Place on Earth,” you’ll go through a new gatekeeper first: facial recognition. The Anaheim resort has expanded facial-recognition technology at entrances to Disneyland Park and Disney California Adventure after months of limited testing, reads Disney’s privacy notice, in which the company states the intention is to make reentry easier and to prevent fraud. Disney says the system is optional: Guests who do not want to use it can enter through non-facial-recognition lanes, where a cast member manually validates their ticket. However, those guests may still have their photos taken, even as Disney says biometric technology is not used on those images. According to Disney’s privacy notice, the system compares a camera image taken at the entrance with the image saved when a guest first used a ticket or pass, converting the images into unique numerical values to look for a match. Disney says it deletes the numbers within 30 days, unless they must be kept for legal or fraud-prevention purposes. And for kids under 18, they can use the system only with parent or guardian consent. Disney didn’t immediately responded to Fortune’s requests for comment. “The security, integrity, and confidentiality of your information are extremely important to us,” reads the company’s privacy notice announcing the technology. “We have implemented technical, administrative, and physical security measures that are designed to protect Guest information from unauthorized access, disclosure, use and modification.” “From time to time, we review our security procedures to consider new technology and methods, as appropriate,” it continued. “Please be aware that, despite our best efforts, no security measures are perfect or impenetrable.” By Friday, the tech was being used in most entrance lines at the two parks, with the Los Angeles Times finding only four lines were spared. Signs near the gates
Apr 28, 2026 · via fortune.com
The suspect in a gruesome double murder that went unsolved for decades died in jail just weeks after he was arrested. Floyd William Parrott, 64, was arrested in Lincoln, Nebraska, on March 25 by officers with the Houston Police Department and the FBI, according to the Harris County District Attorney’s Office. He was charged with capital murder in connection with the killings of 21-year-old Andy Atkinson and 22-year-old Cheryl Henry, who were found dead inside a car parked in a cul-de-sac with injuries to their necks on Aug. 23, 1990, authorities said. The cold case, known as the “Lovers’ Lane Murders,” went unsolved for nearly 36 years. Parrott was found unresponsive in his Nebraska jail cell on Tuesday before his extradition to Texas to face trial, according to the Lancaster County Board of Commissioners. “We ache for Andy’s and Cheryl’s families who were denied their day in court. Our anger for what Parrott took from them is matched only by our determination to keep going,” the DA’s office said in a statement. “His survivors deserve accountability.” ADVERTISEMENT - 1‘Lovers’ Lane Murders’ Suspect Dies in Jail After ArrestUNRESPONSIVEThe suspect in a three-decade-old cold case died before he could face trial. - 2$500M Russian Superyacht Makes Mystery Hormuz Blockade RunCOMING THROUGHNeither Iran nor the U.S. interfered with the mega yacht’s movement. Shop with Scouted This Growth Factor Serum Targets Jowls in Just One WeekLIFT ME UPA next-generation growth factor serum formulated to visibly lift, firm, and smooth—fast and sans irritation.- 3'Stranger Things' Spinoff Renewed for Second SeasonSTAY TUNEDJust last week, it landed in Netflix’s weekly top 10 English-language TV shows. - 4Former NFL Player Dies at 35'DEVASTATING LOSS'The defensive end played in 80 games in the NFL throughout his career. Shop with Scouted Score Up to 70% Off Sex Toys During Lovehoney’s
Apr 28, 2026 · via thedailybeast.com