The courts have been grappling with the use of artificial intelligence (AI) in litigation as its adoption has accelerated in recent years. Practitioners who have filed material containing hallucinated references have been sanctioned by the judiciary, and several jurisdictions have now published formal guidelines regulating the intersection of AI and the law. A 2025 UNSW report found 520 instances of generative AI misuse in legal proceedings across ten jurisdictions between January 2023 and September 2025, including 87 Australian cases, the majority of which involved self-represented litigants. While these cases have led to judicial frustration and court delays, they do not appear to have caused significant personal injustice or irrevocable damage to the parties involved in the legal proceedings. What will happen when AI tools cause real damage? A deepfake is an AI-generated video, image or recording that convincingly emulates a person's appearance or voice. Deepfakes are created using advanced machine learning and make it appear that a person is saying or doing things they never actually did. There are thousands of open-source platforms capable of generating deepfake content. This material can be produced easily, at little to no cost, and at scale. Lawyers and the courts need to consider the impact this may have on the way evidence is treated in litigation and the psychological and procedural burdens it may cause to individuals. When digital evidence can no longer be taken at face value The risk of fabricated evidence is not new. Litigants, practitioners and courts alike have historically encountered allegations of forged signatures, altered documents and manipulated records, and have relied on expert examinations, subpoenas and corroborating evidence to test their authenticity. The challenge today is that AI has made it easier, cheaper and more convincing than ever before. As synthetic content becomes increasingly realistic, distinguishing genuine evidence from
Aug 12, 2026 · via landers.com.au
FORM, the retail execution software company behind GoSpotCheck, has been named to the 2026 Inc. 5000 as it expands its artificial intelligence capabilities and international footprint following its merger with Trax. The recognition comes during a significant year for FORM as it combines GoSpotCheck’s retail execution capabilities with Trax’s image recognition technology. The company serves retailers, consumer packaged goods companies, food and beverage brands and distributors seeking greater visibility into how products are displayed, stocked and executed at the store level. FORM’s growth reflects a broader shift in retail technology toward using AI, computer vision and mobile software to connect corporate merchandising strategies with conditions inside individual stores. For brands managing thousands of products across large retail networks, gaps between planned and actual execution can affect product availability, promotional performance and sales. “2026 has already been a landmark year for FORM,” CEO Ali Moosani said. He pointed to the Trax merger, international expansion and product development combining augmented reality and AI as key developments shaping the company’s strategy. The Trax combination gives FORM a broader technology portfolio at a time when retail execution platforms are moving beyond task management toward automated analysis of physical store environments. Image recognition can help companies identify products and shelf conditions, while mobile workflows give field teams a way to act on that information. FORM’s GoSpotCheck platform uses AI and augmented reality to support frontline retail teams. Its capabilities are designed to improve on-shelf availability, monitor planogram compliance and help field employees identify and address execution problems. The company also offers FORM OpX, while the addition of Trax’s image recognition technology expands its ability to collect and interpret information from physical retail environments. Together, those products position FORM across both the intelligence and execution sides of frontline operations. That combination is increasingly relevant as retailers
Aug 12, 2026 · via citybiz.co
Explore the things
you love
.
Log into Facebook
Email or mobile number
Password
Log in
Forgot password?
Create new account
English (US)
Español
Français (France)
中文(简体)
العربية
Português (Brasil)
Italiano
More languages…
Sign Up
Log In
Messenger
Facebook Lite
Video
Meta Pay
Meta Store
Meta Quest
Ray-Ban Meta
Meta AI
Instagram
Threads
Privacy Policy
Consumer Health Privacy
Privacy Center
About
Create ad
Create Page
Developers
Careers
Cookies
Ad choices
Terms
Help
Contact Uploading & Non-Users
Meta © 2026
Aug 12, 2026 · via facebook.com
About
Press
Copyright
Contact us
Creators
Advertise
Developers
Terms
Privacy
Policy & Safety
How YouTube works
Test new features
NFL Sunday Ticket
© 2026 Google LLC
Aug 11, 2026 · via youtube.com
- DHS has approved Helix surveillance platform for use by Secret Service - USSS conducted a privacy impact assessment to compare privacy risks versus advantages of the platform - DHS has reported a decrease in PIA releases in 2026, despite being a prerequisite for deploying surveillance technology The Department of Homeland Security has approved a new Secret Service surveillance platform even as the agency’s broader pace of privacy impact assessments has slowed significantly this year, FedScoop reported Friday. The system, known as Helix, pulls together video feeds, facial recognition capabilities and license plate reader data drawn from cameras at protected sites and existing image databases. The platform is used by Secret Service personnel to watch for potential threats near sensitive locations, including the White House complex and the area surrounding the Capitol. Don’t miss your chance to sit down with DHS leadership as the department navigates a record 65-percent budget boost and reshapes its artificial intelligence, cyber and border security priorities. Save your seat at the 2026 Homeland Security Summit on November 10, when top agency voices and industry leaders will map out exactly where the next wave of opportunity is headed. Why Was the PIA for the Helix Platform Necessary? In its privacy impact assessment, the Secret Service explained that Helix’s ability to link identifiable information — including video imagery, facial data, biometric templates, and license plate numbers tied to time and location — created a greater privacy risk than any of the underlying data sources posed on their own, making a formal review necessary. A Secret Service spokesperson characterized Helix as part of the agency’s ongoing modernization efforts rather than an entirely new capability, though the agency declined to specify when the platform was first put into use, citing operational security around its protective mission. What Do Privacy
Aug 11, 2026 · via executivegov.com
Large language models guess well and explain poorly. Rule-based systems explain perfectly and struggle with messy real-world data. Neurosymbolic AI tries to get both at once, pairing a neural network’s pattern recognition with a symbolic engine’s hard logic, so a system can read a natural-language question and still get the math exactly right. Google, IBM, Amazon, and a wave of funded start-ups are already shipping it. We walk through how the architecture works, who’s building it, and what the market actually looks like once you cut through wildly different forecasts. Neurosymbolic AI combines two eras of AI research that used to run on separate tracks: neural networks and symbolic AI. Neural networks power today’s large language models and image generators, and they excel at pattern recognition, reading messy data, and producing probabilistic guesses. They also struggle with strict math, hallucinate facts, and operate as black boxes, systems that produce an answer without exposing how they got there. Symbolic AI runs the older playbook: hard-coded rules, knowledge graphs, strict logic. It delivers full accuracy and full transparency and handles math cleanly, and it falls apart on messy real-world data or natural human language. Neurosymbolic systems combine the two directly: A neural network perceives and parses messy input, a natural-language prompt included, then hands the extracted information to a symbolic logic engine that reasons through facts, runs the math, and checks the rules. The combination needs far less training data than a pure neural approach, stays explainable at every step, and holds up against hallucination whenever the task involves verifiable facts or logic. Three architectural patterns show up repeatedly. In the first, a neural model extracts structured information, objects, relations, and tokens, and passes it to a symbolic or logic engine for reasoning; large language models like the GPT family fit this
Aug 11, 2026 · via jonpeddie.com
NEW YORK--Facial recognition and liveness detection have become standard defenses against digital identity fraud, but fraud experts say a new generation of inexpensive artificial intelligence tools is rapidly eroding those protections, allowing organized criminal rings to create synthetic borrowers capable of evading detection for months—or even years. During a recent interview with BankInfoSecurity, Matt Vega, chief fraud strategist at Point Predictive, said one of the most alarming developments is the emergence of "camera injection" kits that allow criminals to bypass facial liveness checks altogether. "Facial liveness checks are the foundation for mobile device verification. But now injection kits selling for about $30 enable fraud rings to hot-wire a mobile device's camera feed or inject a deepfake stream into the verification process, bypassing checks many institutions still rely on," Vega said. Unlike earlier deepfake attacks that attempted to fool a camera by mimicking facial movements, camera injection attacks feed manipulated video directly into a device's camera pipeline, preventing many liveness detection systems from ever seeing a genuine image. According to Point Predictive's Fraud Risk Intelligence Report, these attacks have evolved rapidly from specialized hacking techniques into inexpensive commercial products marketed through Telegram channels and underground forums. Point Predictive cited reporting by MIT Technology Review, which documented dozens of online marketplaces openly selling AI-powered camera injection tools and biometric bypass software, dramatically lowering the technical expertise required to conduct sophisticated identity fraud. Criminals Are Playing The Long Game Vega told BankInfoSecurity that bypassing onboarding is only the beginning of today's fraud schemes. "Once an injection kit clears that first gate, the fraud doesn't stop at onboarding and the identity keeps building," Vega said. "Synthetic profiles typically take six to 18 months to mature, using agentic artificial intelligence to automate payments on secured cards and micro trade lines to build a healthy repayment
Aug 11, 2026 · via cutoday.info
Violent crime rates dropped in major cities this year: Reports WASHINGTON (TNND) — Violent crime rates have dropped in major cities across the U.S. this year, according to reports. The Major Cities Chiefs Association released a report on Sunday that included 66 U.S. law enforcement agencies that saw rates of homicide, rape, robbery, and aggravated assault drop overall compared to last year, The Hill reported. According to the report, homicide rates dropped 17.2%, rape dropped 6.2%, robbery by 16.6%, and aggravated assault by 5.7% from January through June. The MCCA noted that the numbers are preliminary. The Council on Criminal Justice's mid-year update last month reported similar lows, noting a continued decline starting in the second half of 2022, following a peak during the COVID-19 pandemic and Black Lives Matter protests. “There is no single or easy explanation for these historic lows,” the CCJ wrote, but listed "changes in criminal justice operations and strategies, technological advances, and broad shifts in society and culture" as possible options. Last year, U.S. violent crime rates saw the largest decrease since 1937, according to FBI nationwide crime data. Overall, violent crime decreased by about 9.3% in 2025, according to the data. Rape dropped by nearly 8%, and robbery decreased by about 18.5%, the FBI reported. Starting last year, President Donald Trump sent the National Guard and federal agencies to several major cities to crack down on crime. "Under President Trump’s leadership, we’ve been able to help deliver the most historic run of crime reduction in American history – 20% decline in the murder rate, 184% increase in violent gang takedowns, 2,700 kilos of fentanyl seized (up 56% and enough to kill 203 million Americans), and more," FBI Director Kash Patel said in June. The Department of Homeland Security has been implementing new technology to
Aug 11, 2026 · via fox17.com
Mexico’s new banking rules signal a new era for biometric identity verification By Jesús Aragón, CEO and Co-founder of Identy.io Opening a bank account today looks very different than it did a decade ago. Behind each transaction is a growing reliance on advanced identity verification technologies that help protect both customers and banks. As AI-powered fraud becomes more sophisticated, financial institutions are under pressure to strengthen identity verification without adding friction for customers. AI-driven deepfake fraud is projected to increase 495 percent in 2026 compared to 2025, highlighting how quickly fraud techniques are evolving. Rather than isolated incidents, financial institutions are confronting a broader shift in which AI enables fraud to scale more rapidly, making it increasingly difficult for traditional authentication methods to keep pace. In Mexico, that effort has taken the form of new requirements issued by the National Banking and Securities Commission (CNBV). On July 1, 2026, the agency published amendments to the General Provisions Applicable to Credit Institutions, expanding existing identity verification requirements to include facial biometrics alongside fingerprint verification for in-person banking transactions. Previous regulations relied primarily on fingerprint verification for in-person identity confirmation. Under the new rules, financial institutions have 90 days to implement liveness verification methods while continuing to protect biometric information from being sold, shared or transferred to third parties. The amendments also require facial biometric verification to achieve a minimum 90 percent match against government identity records, strengthening identity authentication requirements across financial institutions. The update reflects a trend across Latin America. Governments are placing greater emphasis on the use of biometrics, financial services are becoming more digital and fraud is growing more sophisticated. Organizations such as the FIDO Alliance were founded to help advance stronger authentication standards as identity threats continue to develop. Together, regulations and industry standards reinforce the importance
Aug 11, 2026 · via biometricupdate.com
Over the weekend, two popular gay bars in San Francisco halted the use of ID verification technology after pushback. Badlands and Toad Hall, both in the Castro neighborhood, opted to pause using Patronscan Guard+, an AI assisted identification reader, similar to what you may see at an airport, after community concerns were raised. Watch NBC Bay Area News free wherever you are Patronscan is an ID scanning technology that can identify fake IDs at a forensic level through comparisons with known fake templates and comparisons with government provided security measures. Servall Data Systems, the makers of Patronscan, also offers a flagging service that can help bars and venues alert their employees of problematic customers and share that information with other establishments that utilize this technology. Opponents are concerned that it is data collection of vulnerable populations and yet another form of tech surveillance in the age of license plate readers and even facial scanning AI. Fight for the Future, an advocacy organization that has been working to prevent AI surveillance technology from being utilized in music venues and bars around the country, called for a boycott all bars in San Francisco that use Patronscan, until they discontinue its use. "We have heard the concerns regarding the use of Patronscan at Badlands and Toad Hall in San Francisco, and we are listening," said Brian Aranda, director of operations at both bars, in a statement released on Friday. "Effective immediately, we are pausing the use of Patronscan while we review our ID verification and security practices. The safety of our guests and staff remains a priority, as do privacy and trust. Our security teams will continue to thoroughly check IDs manually during this time." Evan Greer, director of Fight for the Future said that their organization has been working on campaigns to
Aug 11, 2026 · via nbcbayarea.com
Figures Abstract As a fundamental acoustic component, the buzzer is widely used in various electronic systems. The iron core is a critical element in buzzers for supporting the coil, and it is currently fed primarily by mechanical methods. To further improve the automatic feeding efficiency of iron cores, a machine vision-based detection method is proposed to achieve core localization, pose recognition, and notch-angle measurement. The method first employs the Hough transform to locate iron cores on the vibratory tray and exclude overlapping cores. It then statistically counts the edge pixels around each core's center to select only those facing upward. Finally, by traversing the core's circumference, it pinpoints the notch localization and computes its angle. This providing the necessary data support for the automatic grasping and placement of iron cores by the manipulator. Experimental results demonstrate that the Hough transform algorithm adopted in this paper achieves a mean relative localization error of only 2.61%, a recognition precision of 100% for front-up iron cores, and an average notch-angle measurement deviation of 1.06°. Compared with YOLOv8, the proposed method offers clear advantages in both detection accuracy and practicality. Citation: Liu X, Sun C, Wang C, Huang X, Zhu R, Hu C (2026) Machine vision-based detection method for buzzer iron cores. PLoS One 21(8): e0354351. https://doi.org/10.1371/journal.pone.0354351 Editor: Wislei Riuper Osório, UNICAMP, University of Campinas, BRAZIL Received: April 25, 2026; Accepted: July 7, 2026; Published: August 11, 2026 Copyright: © 2026 Liu 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: The data underlying the results of this study are available from GitHub (https://github.com/obito0330/Leoxy). Funding: This research was funded by National Nature Science Foundation of
Aug 11, 2026 · via journals.plos.org
Moscow police beat 15-year-old after AI facial recognition system misidentifies him as drug suspect Moscow police beat a schoolboy after an AI-powered facial recognition system mistakenly flagged him as a criminal suspect, journalist Ksenia Sobchak reported. The incident took place in early May in Moscow’s Lomonosovsky District, the boy’s father said. Two police officers attacked the 15-year-old from behind, knocked him to the ground, and beat him — then took him to a police station and accused him of distributing drugs. Police did not call the boy’s mother until two hours later, telling her that her son was a suspect based on data from an AI system that had allegedly found an 85–90% match with a person named in a wanted notice. After questioning, the mother and son were released. Doctors later diagnosed the teenager with a concussion, a head injury, hematomas, bruises, and abrasions. According to the official police account, the teenager actively resisted arrest, and officers used “combat fighting techniques” against him. The boy’s parents have appealed to the Investigative Committee, Russia’s principal federal investigative agency, but for four months have received only form letters saying a review is underway. At Meduza, we are committed to transparency about our use of artificial intelligence in the newsroom. The story you’re reading was written by one of our living, breathing journalists and translated from Russian using an AI model configured to follow our strict editorial standards. This translation process is the result of extensive testing and refinements to ensure our English-language coverage is timely and accurate. A Meduza editor reviews every draft before publication. If you find any errors in this translation, please contact us at [email protected]. To read Meduza’s exclusive content in English, please subscribe to our newsletter.
Aug 11, 2026 · via meduza.io
Figures Abstract Salient Object Detection (SOD) is concerned with isolating the visually most noticeable objects in an image via precise segmentation. Previous approaches, especially those that adapt the Segment Anything Model (SAM), often yield saliency maps that contain incomplete object masks, blurred boundaries, and a lack of fine-grained details. This issue is especially severe for scenes with objects of different scales or complex textures. We argue that these issues stem from three inherent limitations of existing adaptation strategies: (1) existing adapters rely on rigid multi-scale fusion strategies, lacking learnable cross-scale calibration to handle objects of diverse sizes, (2) simple feature concatenation ignores cross-level semantic correlations, (3) spatial-domain operations inevitably discard high-frequency details. The proposed WMW-SAM, a Weighted Multi-Scale and Wavelet-Enhanced SAM, is designed to handle these limitations for SOD. Specifically, we develop the Weighted Multi-Scale Adapter (WMSA), which utilizes learnable weighting across multiple receptive fields to calibrate features of different scales. Then, our Multi-level Feature Cross-fusion Module (MFCM) employs cascaded top-down cross-attention to facilitate deep interaction that connects high-level semantics with low-level details. Finally, we develop a Detail Enhancement Module (DEM) that leverages the Discrete Wavelet Transform (DWT) to explicitly extract and enhance high-frequency sub-bands. This operation effectively recovers sharp boundaries and intricate textures, which are often overlooked by spatial-domain operations. Extensive experiments on multiple benchmark datasets demonstrate the superior performance of our proposed WMW-SAM, which achieves accurate and detailed saliency predictions. Citation: Liu Z, Tian D (2026) Weighted multi-scale and wavelet-enhanced Segment Anything Model for salient object detection. PLoS One 21(8): e0355742. https://doi.org/10.1371/journal.pone.0355742 Editor: Yongjie Li, University of Electronic Science and Technology of China, CHINA Received: May 4, 2026; Accepted: July 25, 2026; Published: August 11, 2026 Copyright: © 2026 Liu, Tian. This is an open access article distributed under the terms of the Creative Commons Attribution License, which
Aug 11, 2026 · via journals.plos.org
Patronscan ID verification paused at popular gay bars in San Francisco due to privacy concerns San Francisco’s Castro District is rolling back the use of AI‑assisted face scanning technology after LGBTQ+ community members raised alarms over privacy, surveillance and the creation of sensitive data profiles. Two prominent gay bars, Badlands and Toad Hall, said they would stop using Patronscan ID verification following weeks of criticism. “We have heard about the concerns regarding the use of PatronScan at Badlands and Toad Hall in San Francisco, and we are listening,” Toad Hall said on an Instagram post. “Effective immediately, we are pausing the use of Patronscan while we review our ID verification and security practices.” Staff will return to manual ID checks while the review is underway. The scanners were introduced earlier this year but have faced backlash. Patrons say the technology felt intrusive in spaces historically seen as refuges for people who may not be publicly out. The venues use an ID scanner and a camera — Patronscan Guard+ — that captures photos of patrons. These are stored on a database, for up to 21 days, and are shared on a network to which venues can access if they subscribe. Eight venues in San Francisco use the technology, according to SFGate. Patronscan’s system collects a customer’s ZIP code, photograph, date of birth, gender and the expiration date of their ID, according to the company’s website. Company spokesperson Rhiannon Mosoronchon told SFGate that it does not use facial recognition but is stored as a timestamped image of the patron. The company says data is stored for about 21 days to assist law enforcement investigations and is not sold or shared with third parties. It argues its technology complies with California law, which permits ID verification systems to prevent underage drinking and fraud.
Aug 11, 2026 · via biometricupdate.com
Vadzo Imaging validates Real Time Dynamic ROI streaming on Bolt-2020CRS Full 20MP Color MIPI camera for Embedded Vision Applications Vadzo Imaging's Bolt-2020CRS pairs 20MP color resolution with the Onsemi AR2020 sensor's enhanced near infrared response giving security, robotics, and industrial camera products usable color detail from bright daylight through dusk, and also supports Dynamic ROI functionality, allowing embedded vision systems to selectively capture and process important areas within the frame while reducing processing load and optimizing bandwidth utilization. FORT WORTH, Texas, August 11, 2026 (Newswire.com) - Vadzo Imaging, a provider of embedded vision camera products for OEMs and system integrators, today announced the Bolt-2020CRS 20MP Color MIPI Camera built on the Onsemi AR2020 sensor. The new camera module extends color imaging performance into low light and near infrared conditions that typically force a switch to a monochrome sensor or additional visible light illumination. Security integrators, robotics developers, and industrial inspection teams gain a single MIPI CSI-2 camera that holds resolution, color accuracy, and exposure stability from daylight through near dark scenes. The Low Light Color Imaging Problem in Embedded MIPI Camera Design Standard color image sensors lose a large share of incoming photons to the Bayer filter array that separates red, green, and blue light before it reaches each pixel. That loss becomes severe as ambient light drops toward dusk, indoor low light, or nighttime conditions common in security corridors, warehouse aisles, and outdoor inspection stations. Many embedded vision programs respond by switching to a monochrome sensor after dark, which sacrifices color information exactly when identification tasks such as license plate reading, apparel color matching, or intruder clothing description need it most. Adding supplemental visible light illumination is often not possible in covert security placements, wildlife-adjacent industrial sites, or battery-powered robotics platforms where every added watt shortens runtime. Dynamic ROI
Aug 11, 2026 · via newswire.com
Mayflower has become the second Central Arkansas city in the last week to ditch its Flock Safety license plate readers. Interim Police Chief Brittany Byrd announced Monday that after “evaluating the capabilities” of the cameras, Mayflower will discontinue their use and will “rescind their deployment within the city immediately.” The move comes three days after the city of Cabot announced its decision to stop using Flock cameras pending a “comprehensive review.” Last month, the police chief in Pea Ridge announced his city would no longer use the devices. In her announcement, Byrd acknowledged Flock cameras have aided in the investigation of criminal cases across the country, including in Mayflower. But despite the benefits, she wrote, “we believe them to be intrusive in nature.” “The Mayflower Police Department will not participate in tactics which compromise or infringe upon the liberties entitled to law-abiding citizens,” Byrd wrote. Flock license plate reader cameras take pictures of passing cars and log their license plate numbers, comparing the captured information against so-called hot lists assembled by the law enforcement agencies that operate them. Flock Safety, the company that manufactures the devices, states on its website that the cameras don’t use facial recognition software but boast the ability to search captured images based on a description of a vehicle — such as its color, make and model, or unique features such as bumper stickers — if a plate number isn’t known. Many cities in Arkansas use the cameras, including Little Rock, North Little Rock, Conway, Hot Springs, Benton, Bryant, Jacksonville and Sherwood. Nationally, more than 5,000 law enforcement agencies have bought Flock’s license plate readers, the company’s website states. Byrd said Mayflower signed a multi-year agreement with Flock Safety in 2023 that runs through 2027. “Police Department personnel have been communicating with representatives from Flock Safety
Aug 11, 2026 · via nwaonline.com
Courtesy of Gentle MonsterFashion / LongreadHow AI infiltrated fashion, one It Girl at a timeBy making its technology chic and aspirational, the AI industry attempts to win over its sceptics – but is everyone falling for the ruse?ShareLink copied ✔️August 11, 2026FashionLongreadAugust 11, 2026Text Natasha Cornelissen AI no longer wants to limit itself to a techy male consumer base. Increasingly, it’s striving to become a cool-girl staple, with companies seeking to capitalise on the trend-setting potential of celebrities, influencers and the cloutier corners of online culture. From Steven Meisel-shot smart glasses starring Kaia Gerber and Hoyeon Jung to Gentle Monster’s forthcoming AI eyewear fronted by Anok Yai and Alex Consani, fashion and AI are becoming ever more intertwined. The timing is significant. According to recent research, women are more sceptical of AI than men and more likely to think it will have a negative impact on society. For those betting big on AI’s rise, this poses a significant challenge. Despite the gender pay gay gap, women are responsible for up to 85 per cent of household purchasing decisions in the US, possessing a spending power of over $17 trillion. In line with their increasing financial independence and higher levels of disposable income, women are a more important market for luxury brands than ever before. Consider the latest product release from Snapchat’s parent company Snap: augmented reality (AR) glasses, called SPECS. Designed with a sleek aviator frame and made from what Snap calls “plastic titanium”, the $2,195 glasses are positioned as much as a fashion object as a piece of technology. They were launched in June with a campaign shot by legendary fashion photographer Steven Meisel. Cult eyewear designer Gentle Monster has also hopped on the bandwagon, partnering with Google and Samsung on its own AI glasses, with more designs due
Aug 11, 2026 · via dazeddigital.com
Live facial recognition technology is set to be used at selected London Underground stations as part of a new British Transport Police (BTP) trial aimed at identifying people wanted for serious offences. The first deployment will take place at Victoria station on Tuesday, with the technology expected to be used at other key Tube stations during the trial. Transport for London (TfL) said the cameras will compare images of people passing through designated areas with a police watchlist. The system is intended to help officers identify people wanted in connection with high-harm offences, including sexual offences, robbery and knife crime. BTP will decide where and when to deploy the technology using intelligence and analysis of crime data, with the aim of targeting locations where it could have the greatest impact. TfL said that images of people who do not match anyone on the police watchlist will be automatically deleted immediately. The trial will operate within clearly marked areas at stations. TfL said this will allow passengers to make an informed decision about whether to pass through the areas, while BTP said alternative routes will be available. The Tube operation expands on a separate BTP trial involving major railway stations across London that began in February. Between February and July, facial recognition technology was deployed 18 times at major London railway stations, during which more than 530,000 faces were scanned. According to BTP’s published deployment records, none of those scans resulted in a confirmed match or an arrest. The system generated one false identification during the period. Siwan Hayward, director of security, policing and enforcement at TfL, said: “Everyone should be able to travel without fear of harassment, intimidation or violence. “Tackling violence against women and girls and preventing sexual offending on our network remains a key priority. “Live facial recognition
Aug 11, 2026 · via londragazete.com
Abstract Foundation models (FMs) are driving a prominent shift in biomedical imaging, from task-specific models to unified backbone models for diverse tasks. This opens an avenue to integrate imaging, pathology, clinical records and genomics data into a composite system. However, this vision contrasts sharply with modern medicine’s trajectory towards more granular sub-specialization. This tension, coupled with data scarcity, domain heterogeneity and limited interpretability, creates a gap between benchmark success and real-world clinical value. We argue that the immediate role of FMs lies in augmenting, not replacing, clinical expertise. To separate hype from reality, we introduce real-world evaluation and assessment of FMs (REAL-FM), a multi-dimensional framework assessing data, technical readiness, clinical value, workflow integration and responsible artificial intelligence. Using REAL-FM, we find that although FMs excel in pattern recognition they fall short on causal reasoning, domain robustness and safety. Clinical translation is hindered by scarce representative data for model training, unverified generalization beyond over-simplified benchmark settings and a lack of prospective outcome-based validation. This Perspective provides clinicians with a practical way to interpret FM claims, identify where these systems may safely support imaging workflows and recognize why human oversight remains indispensable. For developers, it defines the validation, workflow, safety and governance requirements that must be met before FMs can become clinically reliable tools. We envision that the path forward lies not in a monolithic medical oracle, but in coordinated subspecialist AI systems that are transparent, safe and clinically grounded. This is a preview of subscription content, access via your institution Access options Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription $32.99 / 30 days cancel any time Subscribe to this journal Receive 12 digital issues and online access to articles $119.00 per year only $9.92 per issue Buy this article - Purchase on SpringerLink -
Aug 11, 2026 · via nature.com
Two San Francisco bars recently discontinued their use of Patronscan Guard+, an AI-powered facial recognition and surveillance technology that uses cameras and scanners to collect patrons’ personal data, including an image of their face and government-issued ID as well as their names, addresses, genders, and even patron behavior inside the venue. Patronscan is used to flag “blacklisted” customers, track fake IDs, assist with criminal investigations, and alert other companies using its services about flagged patrons. However, opponents of the tech worry the collected data could help empower anti-LGBTQ+ attacks, especially amid the current administration’s persecution of LGBTQ+ people and their allies as terrorist extremists. Related “Effective immediately, we are pausing the use of Patronscan while we review our ID verification and security practices,” Brian Aranda, director of operations at both Toad Hall and Badlands, told the San Francisco Chronicle. “Our security teams will continue to thoroughly check IDs manually during this time. We appreciate everyone who has shared their concerns and remain committed to providing a safe and welcoming environment for all.” Patronscan’s website says it doesn’t sell or share any patron’s personal information with third-party companies. But, despite its assurances that it doesn’t conduct facial recognition analysis and only stores patron’s data for 21 days (unless the patron requests the company to delete it beforehand), the company itself admitted that it saves the data of flagged patrons for anywhere from one to five years. Never Miss a Beat Subscribe to our newsletter to stay ahead of the latest LGBTQ+ political news and insights. “For bar and nightclub clients, we collect: full name, date of birth, gender (if available on ID), zip code (if available on ID), and expiry date (if available on ID). If an additional authenticity check is required, a venue can run a one-time comparison of the
Aug 11, 2026 · via lgbtqnation.com