The Zooniverse, a NASA grantee that runs the world’s largest platform for online people-powered research, has reached an extraordinary milestone: 1 billion classifications contributed by volunteers around the world. This milestone is a celebration of everyone who has marked a dip in a light curve, confirmed the presence of a moving object in a short video, or identified species in a camera trap image. Each of these small contributions collectively advances our understanding of the universe. A total of 31 NASA-sponsored citizen science projects have been hosted on Zooniverse, accounting for 120 million classifications by 324 thousand volunteers since 2020. Through projects like Planet Hunters TESS, Daily Minor Planet, Backyard Worlds: Planet 9, Space Umbrella, and Snapshot Wisconsin, volunteers help discover exoplanets, identify near-Earth objects and asteroids, search for brown dwarfs and planetary systems, analyze effects of the solar wind, and inform wildlife management decisions. These projects have led to 96 scientific publications, and 56 of these articles feature NASA citizen scientists as co-authors to recognize the significance of their research contributions. These efforts demonstrate how public participation can accelerate discovery by combining human curiosity and pattern recognition with data from NASA missions and observatories. Collaboration between volunteers, scientists, and computing technology will be even more important in the future as we tackle enormous and complex datasets, like those from NASA’s upcoming Nancy Grace Roman Space Telescope. "One billion classifications represent far more than a number; it's one billion moments of curiosity transformed into meaningful contributions to research," said Laura Trouille, principal investigator of Zooniverse and vice president of Science Engagement at the Adler Planetarium. "Every classification on Zooniverse brings us one step closer to new discoveries and a deeper understanding of our universe, our world, and ourselves.” Zooniverse is the world's largest platform for people-powered research. Co-founded by the
Jul 10, 2026 · via science.nasa.gov
Theatreland and Soho are two must-tick-off stops for tourists looking to get the full London experience — as well as being places that locals will occasionally visit for a night out. But they’re also crime hotspots: the West End and Soho remain among London’s most crime-hit areas, with theft, robbery and violence among the offences most commonly recorded. To try and tackle that scourge, the Metropolitan Police plans to expand a trial the force says was successful in tackling crime in Croydon into more densely populated areas. Some 173 arrests were made after suspects were identified using live facial recognition (LFR) technology in Croydon, the Met says, with only one face out of 470,000 scanned during a six-month pilot being incorrectly identified. The Met has pointed to individual Croydon cases that have since resulted in prison sentences, but it has not published a full breakdown for all 173 arrests made during the six-month pilot. “We want to build on our success by introducing this capability to the West End and Soho by December,” said Sir Mark Rowley, the Met commissioner. “The use of static cameras will help us continue cutting crime in high-footfall areas in central London.” “If you haven’t been a society in which you don’t know if you’re being watched at any given time, there’s a chilling effect” Jasleen Chaggar, senior legal and policy officer at civil liberties group Big Brother Watch The goal is to put it in place in time for the pre-Christmas period. But what impact will it have on everyday life in these areas? “If you haven’t been a society in which you don’t know if you’re being watched at any given time, there’s a chilling effect,” says Jasleen Chaggar, senior legal and policy officer at civil liberties group Big Brother Watch. Live facial
Jul 10, 2026 · via standard.co.uk
Vadzo Imaging Positions AR0521 Liveness Detection USB Camera for Transactional Kiosk Face Verification: Low-Noise 5MP Imaging with NIR Compatibility for Anti-Spoofing Vadzo's Falcon-521CRS is a 5MP color USB camera built on the ON Semiconductor AR0521 sensor, delivering low read noise imaging with NIR spectral response and USB 3.2 Gen 1 connectivity, designed for embedded face verification, anti-spoofing pipelines, and biometric authentication terminals requiring a compact UVC-compliant module without custom driver development. FORT WORTH, Texas, July 10, 2026 (Newswire.com) - Vadzo Imaging today announces the availability of the Falcon-521CRS, an AR0521 liveness detection USB camera built for deployment in transactional kiosks, eKYC stations, access control terminals, and biometric authentication systems. Powered by the ON Semiconductor ARQ0521 sensor and delivered over USB 3.2 Gen 1 with full UVC compliance, the Falcon-521CRS combines 5MP resolution with low read noise characteristics and near-infrared band sensitivity in a compact board-level module ready for direct OEM integration. The Imaging Challenge in Biometric Kiosk and Face Verification Systems Deploying a reliable liveness detection camera inside a transactional kiosk or access control terminal involves imaging constraints that general-purpose USB modules are not engineered to meet. Face verification of workflows depends on consistent high-fidelity capture across a range of ambient lighting conditions, from controlled indoor environments to high-glare outdoor kiosk installations and low-light lobby areas. A sensor with poor noise performance delivers image data that undermines feature extraction accuracy in face recognition algorithms, producing elevated false rejection rates and degraded liveness classification scores. Anti-spoofing pipelines introduce a second requirement: NIR response. Presentation attack detection frameworks using near-infrared illumination at 850 nm or 940 nm require a sensor that captures a usable signal in these bands without a monochrome-only optical path. Color sensors without NIR sensitivity cannot participate in these pipelines without additional hardware overhead. For engineers building kiosks and
Jul 10, 2026 · via newswire.com
OPINION article Front. Artif. Intell. Sec. Machine Learning and Artificial Intelligence AI Without Representation Is Just Inequity at Scale: On the Exportation of Unrepresentative Artificial Intelligence Models to the Global South - CACésar Abelardo Tinco Aliaga 1 - VGVasco Gerardo Hinostroza Fuentes 2 - HAHezerul Abdul Karim 3 - NANouar AlDahoul 4 - MJMyles Joshua Toledo Tan 5,6,7,8 - 1. Department of Electrical and Mechatronics Engineering, University of Engineering and Technology, Lima, Peru - 2. Department of Computer and Information Science and Engineering, Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL, United States - 3. Centre for Image and Vision Computing, Centre of Excellence for Artificial Intelligence, Faculty of Artificial Intelligence and Engineering, Multimedia University, Cyberjaya, Malaysia - 4. Department of Computer Science, Division of Science, New York University Abu Dhabi, Abu Dhabi, United Arab Emirates - 5. Department of Electrical and Computer Engineering, Herbert Wertheim College of Engineering, University of Florida, Gainesville, FL, United States - 6. Department of Epidemiology, College of Public Health and Health Professions and College of Medicine, University of Florida, Gainesville, FL, United States - 7. Biology Program, College of Arts and Sciences, University of St. La Salle, Bacolod, Philippines - 8. Department of Natural Sciences, College of Arts and Sciences, University of St. La Salle, Bacolod, Philippines Select one of your emails You have multiple emails registered with Frontiers: Notify me on publication Please enter your email address: If you already have an account, please login You don't have a Frontiers account ? You can register here Abstract A common assumption in discussions of artificial intelligence (AI) is that increasing demographic representation in training data is sufficient to mitigate bias. Under this view, failures in facial recognition, clinical classification, and language generation are treated primarily as problems of coverage that can be
Jul 10, 2026 · via frontiersin.org
Dutch watchdog finds most smartphones can be unlocked with just a picture of the owner Most smartphones tested by the Dutch consumer organization Consumentenbond since June 2023 failed a facial recognition security test. Researchers were able to unlock the devices by holding a photo of the owner in front of the camera, Consumentenbond reports. The Consumentenbond found that 92 of the 133 tested smartphones still available for purchase could be fooled by a photo. That represents 69 percent of all tested devices. The result is worse than three years ago, when researchers were able to unlock 43 percent of tested phones using a photo. Many of the affected devices were made by Motorola, with 19 models failing the test. Oppo had 18 failed models, and Redmi had 16. The problem was not limited to cheaper phones. Several expensive smartphones also failed the test, including the 1,300-euro Oppo Find X Pro, the 1,250-euro Samsung S25 Ultra, and the 750-euro OnePlus 15. Only Apple and Google had secure facial unlocking on all tested devices, according to the Consumentenbond. The consumer group said the results are concerning because buyers cannot assume their phone’s facial recognition system is secure. Many phones with unsafe facial recognition do warn users that the feature is less secure than a password or PIN code. “Smartphones should either have good facial recognition or not have it at all,” Sandra Molenaar, director of the Consumentenbond, said.
Jul 10, 2026 · via nltimes.nl
A planned visit by police facial recognition vans to an Oxfordshire town has been cancelled. Thames Valley Police was set to deploy the technology in Witney Market Square today (July 10), but has now announced it has been cancelled. A spokesperson said: "Our Live Facial Recognition vans were due to be at Market Square, Witney, today (July 10). READ MORE: Hundreds of Oxfordshire villagers without power after 'huge explosion' "This deployment has been cancelled for the time being, due to the team being committed with other demands today. "We will keep you updated on any future deployments." Police had said they would be using the live facial recognition vans in Witney to combat crime. The specialist LFR team was set to be working alongside local officers to identify known suspects, deter crime and help keep our communities safer for everyone. The technology matches a digital image of human faces against known faces in a database and generates possible matches. READ MORE: Multiple arrests in Oxfordshire towns after 'armed robbery' probe The force had said, prior to the event being cancelled, that it would be happy to speak to anyone about how the technology works and how it is used responsibly. The spokesperson said: "If you have any questions or would like to see how the technology works, feel free to come over and chat with our team. "We’re always happy to show you around and explain how LFR is used fairly and responsibly."
Jul 10, 2026 · via oxfordmail.co.uk
Kazakhstan introduces new biometric authentication rules Kazakhstan's Agency for Regulation and Development of the Financial Market and the National Bank have approved new rules governing biometric authentication by financial institutions, Qazinform News Agency reports. The new rules outline how banks and other financial institutions will verify customers' identities using biometric authentication. They also specify when the National Biometric Authentication System must be used and biometric data must be submitted through the National Bank's Identification Data Exchange Center (IDEC). Under the new rules, customers will be authenticated through facial recognition. The system captures their image, matches it against a reference image from a government database or another authorized source, and verifies their identity. To guard against identity fraud involving photographs, videos or other spoofing techniques, the new rules require mandatory liveness detection. During the authentication process, users must respond to at least three randomly generated prompts while the system analyzes their actions in real time. If the verification fails, the procedure is repeated, with authentication deemed unsuccessful after three failed attempts. The regulations also set out requirements for protecting biometric data. Financial institutions must ensure that biometric information is collected, stored and processed in line with established procedures and safeguarded against unauthorized access or disclosure. All stages of the data lifecycle, from collection to destruction, must also be documented. Earlier, Qazinform News Agency reported that Kazakhstan may launch an aerotaxi service between Alatau and Almaty in 2027.
Jul 10, 2026 · via qazinform.com
Hyderabad: Research on making 3D models lighter, improving how artificial intelligence understands videos, and preventing unauthorised 3D reconstruction from images featured among the papers presented by researchers from the International Institute of Information Technology, Hyderabad (IIIT-H) at the Computer Vision and Pattern Recognition (CVPR) 2026 conference and its workshops. Among the award-winning papers, Kunal Bhosikarâs âFast and Robust Mesh Simplification for Generated and Real World 3-D Assetsâ was named Best Paper Runner-Up at a workshop on 3D geometry generation. The paper addresses a key challenge in 3D AI, where highly detailed models are often too heavy to store, render and process efficiently. The work proposes a mesh simplification technique that removes unnecessary triangles while preserving the modelâs shape, curvature and texture, making 3D assets easier to use in areas such as medical imaging and virtual reality. âThe result is lighter 3D models that remain visually accurate but can be processed much faster,â Kunal said. Kunal also presented âPatchPoison: Poisoning Multi-View Datasets to Degrade 3D Reconstructionâ, which explores a way to stop images from being turned into 3D models without consent. The paper introduces a nearly invisible digital patch that can be embedded into images and disrupt AI systems used for 3D reconstruction, producing blurred or distorted outputs. The work points to a possible privacy tool for photographers, creators and businesses seeking to prevent unauthorised reuse of their images. Another IIIT-H paper to win honours was Darshan Singhâs âSRL-CLIP: Efficient CLIP Video Adaptation via Structured Semantic Role Labelsâ, which won Best Paper at a workshop on data-efficient video intelligence. The paper focuses on improving video understanding without relying on massive datasets. Instead of training AI on simple captions, the team used structured semantic labels that capture who is doing what, and in what context. âWith this, we were able to train
Jul 10, 2026 · via timesofindia.indiatimes.com
Facial recognition technology in shops will soon alert police in real time to the presence of serious offenders, with civil liberties groups warning of a “dangerous escalation” towards surveillance and criminalisation in the retail sector. Facewatch, a facial recognition system used by more than 100 businesses including Sainsbury’s, B&M and Spar to monitor thieves, said it was launching a UK-first feature to “alert police instantly when the most serious offenders trigger a live facial recognition match”. Facewatch’s chief executive, Nick Fisher, said the “unique technical development” would be launched in autumn and would warn police in an average of four seconds when the “worst offenders” were flagged on its network. Civil liberties groups have voiced alarm at the development, saying it had “shot on far ahead of the regulation” and was “upending” the way retail crime was dealt with. Charlie Whelton, the policy and campaigns officer at Liberty, said it was concerned about this “untested, opaque development” and the way facial recognition technology had been allowed to “proliferate without anything to govern it”. “It’s not against the law to walk into a shop even if you’ve committed crimes in the past,” he said. “The idea of calling the police on somebody who hasn’t committed a crime, but there’s a concern they might, is really upending the way we do things. And of course, it’s not infallible. These systems do make mistakes, and it’s very hard to argue with that when it happens to you.” A number of people have been forced to leave shops after being falsely identified by Facewatch technology as a shoplifter, with some describing it as “Orwellian” and saying they felt as though they were “guilty until proven innocent”. Evidence suggests black and Asian people are more likely to be incorrectly identified than white people. Britain’s biometrics
Jul 10, 2026 · via theguardian.com
Top photo caption: Chloe Berwick is shown during the Catholic Media Conference in Atlantic City where was named a winner of the inaugural Carlo Acutis Scholarship. Courtesy photos By David Karas, Correspondent Growing up, Chloe Berwick was active in her home parish of St. William the Abbot, Howell – attending religious education classes, receiving the Sacraments, serving as a catechist and co-leading a high school youth group with her twin sister. Berwick also found a way to merge her faith with her passion for writing – and her interest in Catholic journalism recently led to her recognition as a winner of the inaugural Carlo Acutis Scholarship. The scholarship offers full-time college students or recent graduates the opportunity to attend the Catholic Media Conference, the flagship event of the Catholic Media Association, where they can develop their skills, build professional relationships and learn directly from experienced Catholic media professionals. Applicants must submit a letter of introduction, resume, four content samples, an academic transcript, two letters of recommendation and a description of three story ideas for Catholic media. Recognized during the conference in Atlantic City June 16-19, each scholarship provided full conference registration, four nights at the conference hotel and travel reimbursement of up to $500. The award is made possible through generosity of the Catholic Journalist Scholarship Fund, eCatholic and the Catholic Extension Society. “The connections made and lessons learned certainly extend beyond my time at the conference,” Berwick acknowledged. Poised for Success “I am extremely grateful for this experience, especially for the chance to meet my fellow recipients,” Berwick said. “I learned from professionals on a variety of topics, including how to build a brand, engage with influencers, explore AEO (Answer Engine Optimization) analytics, and use visual storytelling and photojournalism techniques. I have already started applying the insights and skills
Jul 10, 2026 · via trentonmonitor.com
Unreliable Facial Recognition Technology Will Soon Be Used by Police Nationwide The Western Australia Police Force announced on 19 June 2026 that it is trialling the use of live facial recognition technology to randomly scan people passing through public areas to potentially match them in real time to a watchlist containing facial images of people law enforcement is looking for. And this quiet little development likely presages continentwide use of the technology. WA police are trialling live facial recognition using vans deployed to public places. The cameras are attached to the vehicle or a stand beside it. The van states that police are using live facial recognition and there is additional signage notifying passersby. Officers with computer equipment are inside the van. And the watchlist contains 4,000 photos: some of suspects, while others are missing persons. Ten van deployments took place across a number of Boorloo-Perth districts over the first seven days of the trial. WAPOL states that it scanned 131,478 civilians, which generated 33 alerts, one of which was a false positive detection. Eighteen arrests had taken place, 21 sex offenders were identified and engaged with, while two welfare checks took place and an assault incident in a mall was foiled. The trialling of live facial recognition in WA is rather unobtrusive. It will garner much less national attention than if New South Wales police attempted it. This method of introduction also avoids the pushback that the now defunct 2017 Turnbull government proposal to establish a national facial recognition system, which sought to match civilian ID photos with CCTV images in real time, received. WA police commissioner Col Blanch last month explained, “This is the first time in Australia that we will be using this technology. But it is used in other jurisdictions, particularly in the UK.” Yet, the
Jul 10, 2026 · via sydneycriminallawyers.com.au
On Tuesday, Meta launched “Muse Image,” a new AI image-generation feature that allows users to create original images, edit existing photos, and even generate custom ads directly within its apps. But one capability has quickly become the center of controversy. Muse Image allows users to generate AI images using photos from public Instagram accounts. As long as a person’s profile is public, another user can tag that account and use their images as part of an AI-generated creation. (Only private accounts and accounts belonging to users under 18 are automatically excluded from the feature.) One huge concern is consent. Users may have no idea that their public photos can be incorporated into AI-generated images by strangers, and they aren’t even notified when someone reuses their public content. Plus, making it easy to manipulate people’s images opens the door to misuse, harassment, impersonation, and nonconsensual image editing. If you’re looking to opt out of this, here’s how you can do it. How to opt out of Meta’s Muse Image generator - Head to your profile and click the three horizontal lines in the top-right corner. - Scroll down to “Sharing and reuse.” - Look for the option that says, “Allow people to use your content on Instagram with AI features on Meta” - Toggle the setting off for both posts and reels. Muse Image arrives at a time when AI tools are being increasingly integrated into social media platforms. As tech companies race to roll out new generative AI features, many experts argue that stronger privacy protections and greater transparency are needed, so users fully understand how their photos and personal data are being used. Public skepticism around AI is already high. According to a Pew Research Center survey, 35% of respondents said they’re more concerned than excited about the growing
Jul 9, 2026 · via techcrunch.com
The agentic caller always rings at scale: Reality Defender explores new AI voice threat Deepfakes remain a potent part of the fraud arsenal, but the industry’s alarm bells have begun ringing over agentic AI. A new article from Reality Defender looks at how an influx of AI customer service agents – or “agentic AI callers” – will transform how contact centers work. Reality Defender VP of Human Engagement Gabe Regan cites numbers from Gartner, which predict that, by 2029, AI will autonomously resolve 80 percent of common customer service issues without human intervention. “Many of those interactions will originate from AI systems calling in, not humans,” Regan writes. “Contact center infrastructure that engineers built for human callers with occasional bot probes is not prepared for autonomous AI agents at scale, and the gap between what current detection catches and what developers design agentic callers to avoid is where the problem lives.” His post provides a detailed breakdown of the agentic AI caller problem, which levels up on the robocall by listening to interactive voice response (IVR) prompts, selecting the correct menu options, and conducting “a full adaptive voice conversation with a human agent if the IVR routes it through.” AI generated deepfake audio mimics human cadence and tone, and can adapt to unexpected questions. “Nothing in the interaction pattern flags the call as non-human, because the creators specifically designed the system generating the call to avoid triggering those flags.” That’s a problem when customer service becomes a purely agentic transaction, in which some agents are calling support centers on behalf of legitimate AI systems. Cracking down on agentic activity risks denying actual customer requests. “Whether the agentic caller is acting on a customer’s behalf or probing for fraud, the contact center needs to know what it is dealing with before
Jul 9, 2026 · via biometricupdate.com
In a bid to lower its GPU costs amid an unprecedented component shortage, Meta is on track to start making the latest versions of its AI-specific chip in September, Reuters reported, citing an internal memo. At least one chip sailed through its testing phase in about six weeks, the memo said. Meta is working with Broadcom on the chip design, but it will use Taiwan Semiconductor Manufacturing Company (TSMC) to manufacture them. It is also buying RAM from Samsung, storage from Sandisk, and fiber-optic equipment from Sumitomo Electric, according to the report. Meta detailed the four new chips, developed under its Meta Training and Inference Accelerator (MTIA) program, in March, some of which are currently in deployment or will be this year or next. The company is taking a modular approach to designing these chips, anticipating that their needs will change as AI evolves rapidly by the time the chips are in production. “Each MTIA generation builds on the last, using modular chiplets, incorporating the latest AI workload insights and hardware technologies, and deploying on a shorter cadence,” the company wrote at the time. The chips are expected to help the company save on buying GPUs from chipmakers like Nvidia and AMD, although it still expects to spend plenty with those providers as well, Reuters reports. Meta intends to use the MTIA chips for training models for its ranking and recommendation algorithms, broader AI workloads, and inference aimed at its applications. The social media company has been producing its own AI chips since 2023. Meta has been spending massively on securing enough compute capacity to power its various AI efforts. The company in April said it expects capital expenditures between $125 billion and $145 billion this year, a lot of which is going toward its AI efforts. The company has
Jul 9, 2026 · via techcrunch.com
Figures Abstract Automated image captioning remains a challenge, as it enables machines to generate context-aware textual descriptions of visual content. Traditional deep learning approaches often rely on lexical overlap and fail to capture semantic relationships among objects, leading to captions that lack contextual richness. This study proposes an encoder–decoder framework that integrates YOLOv5 with a generative transformer to generate descriptive image captions. The proposed model was evaluated against two baselines: CNN-LSTM (M1) and a BERT-based transformer model (M2). M1 achieves BLEU-1 (0.45) and ROUGE-L (0.42) but demonstrates limited semantic understanding with METEOR (0.18) and SPICE (0.07). M2 improves with higher METEOR (0.24) and CIDEr (0.62), although its BLEU scores remain low. The proposed model achieves the highest CIDEr (1.10) and SPICE (0.25), reflecting superior semantic understanding and better capture of object relationships. Despite a lower BLEU (0.40), it significantly outperforms traditional methods in caption quality. To further validate these results, we conducted an expert-based evaluation to assess semantic accuracy, visual grounding, and caption usefulness. The proposed model achieved 93% accuracy in expert evaluations across 500 images, indicating strong contextual alignment with human interpretation. Additionally, we employed exploratory data analysis to examine and visualize the text captions, aiming to gain a deeper understanding of the optimal caption. Citation: Danyal M, Roman M, Shahid A, Yahya M (2026) Contextual image caption creation using object positional embedding and generative models. PLoS One 21(7): e0353466. https://doi.org/10.1371/journal.pone.0353466 Editor: Muhammad Bilal, King Abdulaziz University Faculty of Engineering, SAUDI ARABIA Received: August 26, 2025; Accepted: June 22, 2026; Published: July 9, 2026 Copyright: © 2026 Danyal et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: All data required
Jul 9, 2026 · via journals.plos.org
| When you purchase through links on our site, we may earn an affiliate commission. Here's how it works. | Here is the corrected version of your article. I have fixed the typos, grammatical slip-ups, and punctuation errors while carefully preserving your exact writing style, cynical tone, and conversational flow. Sony has officially announced the RX10 V, a fixed-lens 1″ sensor camera. The new RX10 V sports a 24-600mm-equivalent Zeiss Vario-Sonnar T* f/2.4-4 Zoom lens, which is a bit of a competitor to the Nikon P1100—though that camera will hit the equivalent of 3000mm, which is ridiculous, but people buy them. The Nikon runs about $1200 USD, while the Sony is an eye-watering $2300 USD. I'm not sure who the customer is that will spend that much on a 1″ sensor camera with a mediocre lens, but what do I know? The Panasonic Lumix FZ80D will bring you to a 1200mm equivalent and for under $500 USD. The image quality may not be as good as the Sony RX10 V, but from what I have seen, the latter isn't overly impressive either. This is Sony's second unimpressive, overpriced fixed-lens camera, following the absurdly overpriced, recycled mess that is the Sony RX1R III. After spending a week with that camera, it was clear that Sony is just throwing darts on higher-margin cameras and filling perceived voids in the industry. I'm not even going to put an affiliate link in this article for pre-orders. I'm not a fan of “0-Day” reviews, but none of them seem to be positive. It's almost like people were searching for things to like instead of things to dislike. Here's hoping Canon doesn't go this route with future PowerShot cameras. We need some new and top-shelf fixed-lens cameras, and Canon seems to be the last great hope.
Jul 9, 2026 · via canonrumors.com
For years, Google Photos was the easiest part of my online life. Every picture I took automatically landed in the cloud, faces got tagged without me lifting a finger, and a quick search for "beach" or "dog" brought up exactly what I wanted. But the convenience came with a steep hidden cost: my entire photo library, including thousands of private family images, was sitting on someone else's servers. That made the trade-off clear to me: I was handing over a massive personal history to a tech giant that has every incentive to analyze it, monetize it, or change the terms of the deal whenever it wants. That trade-off finally stopped feeling worth it, so I left Google Photos behind. I've since tried a few different homes for my library, including Sync.com's encrypted cloud storage. This time, I wanted to go a step further: no third party at all, encrypted or not. So I moved my library to Immich. What is Immich, and how does it replace Google Photos? Core features that replace Google Photos' best tools Immich is a self-hosted photo and video management platform built to feel like a drop-in replacement for Google Photos or Apple's iCloud Photos. You run it on your own machine (a server, a NAS, a spare PC, or, in my case, a laptop) and install a companion app on your phone. The app automatically backs up new photos and videos to your server over Wi-Fi, just like the Google Photos app backs them up to the cloud. Once your images land on the server, Immich handles the heavy lifting you're used to: - Facial recognition, which clusters photos by the people in them, so you can browse by face rather than by date - Object and scene detection, so you can search for "sunset"
Jul 9, 2026 · via makeuseof.com
Before OpenAI’s GPT-3 ushered in the era of foundation models, companies built specialized natural language processing models from scratch, training each on large amounts of task-specific data. Today, most organizations start with a general-purpose model like OpenAI’s GPT series, Claude, or Llama and then fine-tune or prompt it to solve their specific needs. Pim de Witte, CEO of General Intuition, thinks embodied AI will follow a similar pattern. Rather than collecting huge real-world datasets to build specialized robot models, he argues the industry should focus on better quality datasets that can produce foundation models capable of transferring intuition about movement and interaction across many environments. “A lot of companies right now are doing lots of specialized work focused on individual embodiments, individual environments, and individual robots,” de Witte told TechCrunch on a recent episode of Equity. Much of that work will become redundant soon, he argues, with the emergence of general models like the one General Intuition has been developing and deploying. “The generalization of the model itself is the product,” he said. “The fact that it has a base level of reasoning about space and time is going to be the reason why people stop collecting hundreds of thousands or millions of hours of real-world data. Because the reality is, you only need a few minutes.” General Intuition built its own such foundation model after training on millions of hours of video game data, including information like what buttons on a controller a human pushed and when. Both de Witte and General Intuition’s lead investor, Vinod Khosla, argue the action data is the key to developing a human-like intuition for spatial-temporal reasoning. The startup last month raised $320 million at a $2.3 billion valuation on the back of that thesis. The company has demonstrated that its current model is
Jul 9, 2026 · via techcrunch.com
RecFaces Updates Id-Time with Integrated Workforce Management and Enhanced Facial Recognition Capabilities RecFaces has announced a major update to Id-Time, its facial recognition solution for workforce time management, attendance control and operational visibility across enterprise facilities. According to the company, the latest release expands Id-Time into a workforce time management and monitoring platform designed for organizations with complex structures, distributed locations, shift-based operations, high volumes of people and requirements for automation and accuracy. The update introduces a biometric time and attendance module built directly into the platform. Organizations can manage work schedules, timesheets, attendance reports, overtime, undertime, breaks, absences, violations and automated notifications through a single interface. The company said the functionality gives HR teams, security personnel, facility managers and executives a unified view of employee attendance and workforce discipline. RecFaces also upgraded the facial recognition algorithm powering Id-Time. The company said the new version uses an algorithm ranked among the top five globally by NIST and improves identification accuracy and consistency in environments with poor lighting, heavy foot traffic or partially visible faces. The release also adds real-time monitoring and diagnostics dashboards for security teams, HR departments and managers. The dashboards provide visibility into system load, terminal status, camera performance, database diagnostics and the number of people currently on site. RecFaces said the feature can also support emergency response efforts by helping security teams determine how many people are located in a specific area. Reporting capabilities have also been expanded with more than 10 report types, including attendance summaries, worked time, overtime, undertime, violations, visitor activity and presence monitoring. Reports can be exported in Excel, CSV and PDF formats for payroll processing, HR analysis or API integration with external systems. The updated platform also introduces visitor flow monitoring, allowing organizations to track visitors separately from employees. The feature records
Jul 9, 2026 · via securityinfowatch.com
Why It Matters The House Foreign Affairs Committee is holding a hearing on the Fiscal Year 2027 Bureau of Industry and Security (BIS) budget Tuesday, July 14 that brings into focus a critical shift in how the U.S. government allocates resources for technology competition and export controls. The Commerce Department's Bureau of Industry and Security faces mounting pressure to fund artificial intelligence (AI) priorities alongside its traditional role managing technology exports and sanctions. At stake is whether federal agencies can adequately resource the AI arms race while maintaining oversight of sensitive technologies flowing to strategic competitors. The hearing reflects intensifying corporate lobbying around defense and AI funding. Several companies filed first-quarter lobbying disclosures targeting FY27 defense appropriations and the National Defense Authorization Act, with AI capabilities and military applications dominating their focus. Companies ranging from autonomous systems specialists to facial recognition firms are positioning themselves for federal contracts and funding. The Big Picture The intensity of AI-focused lobbying surrounding this budget cycle underscores the stakes. Virtualitics Inc. spent $100,000 in quarter 1 2026 on in-house lobbying regarding AI for military readiness and the FY27 National Defense Authorization Act. Hermeus Corp. invested $135,000 in the same quarter seeking funding for advanced aircraft development in the FY27 defense appropriations bill. Other firms targeted narrower AI priorities. ZeroEyes Inc. lobbied on AI detection capabilities for defense appropriations. Shield AI Inc. focused on autonomous aerial platforms and FY27 Defense Appropriations. Ultra Intelligence & Communications spent $75,000 advocating for AI and machine learning initiatives across the federal legislative cycle. The lobbying activity reflects a broader shift toward embedding AI across defense and intelligence operations. Clearview AI Inc. lobbied on facial recognition technology issues across multiple FY27 appropriations bills. Edgerunner AI Inc., Swarmbotics AI Inc., and Unstructured Technologies Inc. each filed separate Q1 2026 disclosures targeting
Jul 9, 2026 · via legis1.com