About Press Copyright Contact us Creators Advertise Developers Terms Privacy Policy & Safety How YouTube works Test new features NFL Sunday Ticket © 2026 Google LLC
San Francisco Gay Bars Are Scanning Patrons’ Faces and Collecting Their Data Several bars in the Castro are now using facial recognition technology that records a person’s name, address, and even their gender. Featured image: cottonbro studio/Pexels Some bars in San Francisco’s famous Castro neighborhood have installed facial recognition kiosks, which collect information from people’s addresses to their genders. According to the Gazetteer SF, at least three bars in the historically LGBTQ+ district use something called a Patronscan Guard+, which is a device by a Canadian company that collects biometric and personal data. The devices collect names, addresses, genders, and patron behavior, the outlet reports. A reporter with Gazetteer SF went to the bars after none of the establishments responded to requests for comment, noting where the device was placed at one of them. “Like most private surveillance cameras, the Patronscan kiosk at Mix hides in plain sight. In the dim light of the bar, the black machine is easy to miss. I was also not instructed to face the camera when I handed my ID to the bouncer; when I asked if I would be photographed, the bouncer told me the camera had in fact already taken my picture,” Cydney Hayes wrote. “They said Mix bouncers are not required to verbally tell each patron that they’re being photographed by the Patronscan device. Instead, they rely on a small informational plaque posted to the kiosk below eye level to inform customers what data is being collected and how it will be used.” A bouncer told Gazetteer SF that the information gathered from the Patronscan device keeps the data for about a month and then deletes it. If a person has been unruly at the bar, that information is saved. The devices have caused some bar patrons to be concerned. “I
AI editing at Apple and an indicator light for glasses – Photo News of the Week Apple adds AI tools to Photos. Meta deletes facial recognition from Ray-Ban glasses. US law demands recording light for smart glasses. WWDC 2026 brought quite a few interesting things for photographers: Apple is equipping its Photos app with three new AI-powered editing tools. While the company is still keeping details about the specific functions under wraps, the integration into OS 27 and “Apple Intelligence” at least promises that the editing will take place directly on the device – a plus for privacy-conscious users. However, whether the tools can do more than the already established AI functions of the competition will only become clear in the fall when OS 27 is rolled out. Meta secretly wipes away evidence What's happening at Meta is much more explosive: After the company vehemently denied integrating facial recognition technology into its Ray-Ban Smart Glasses, developers discovered corresponding code in the app. Shortly thereafter, this code, however, disappeared silently and without official explanation. This not only smells of a guilty conscience but also raises the question of what Meta actually intends to do with the recordings from its glasses wearers. The whole operation is reminiscent of a burglar quickly wiping away fingerprints when leaving the crime scene. Only here, millions of users are potentially affected. Meta continues to claim that it does not use facial recognition – but why the code then? And why the secret deletion? Building trust definitely works differently. Lawmakers demand visible warning lights Speaking of trust: In the US, representatives have introduced a bill that would require smart glasses with recording functions to have a clearly visible warning light. The “Smart Glasses Recording Act” is intended to prevent people from being secretly filmed – a problem
About Press Copyright Contact us Creators Advertise Developers Terms Privacy Policy & Safety How YouTube works Test new features NFL Sunday Ticket © 2026 Google LLC
JACKSONVILLE, Fla. — A southwest Florida man, who Action News Jax first reported was wrongly arrested for trying to lure a child in Jacksonville Beach, is now suing the police department and the Jacksonville Sheriff’s Office for using an AI-based facial recognition system to make the arrest. We first told you about Robert Dillon’s botched 2024 arrest last July. “He just told me that I was wanted out of Jacksonville for a warrant. And when we got in the car, he told me that it was for attempted alluring of a child. And of course I’m completely dumbfounded,” Dillon said. >>> STREAM ACTION NEWS JAX LIVE <<< Since then, we’ve documented one other wrongful arrest case here in Duval using the same technology. Jalil Richardson, from North Carolina, sat in jail here for weeks for a stolen car case. “There was no proper investigation done, um, to even reach out to me or to see if I was even in Florida,” Richardson said in a previous interview with Action News Jax. Richardson was set free after his attorney provided timesheets to show he was at work when the crime was committed. Action News Jax spoke with Dillon and his attorney about their call for a change in policy when it comes to facial recognition. He said he had never been to Jacksonville. “The only time I’ve ever even been close to Jacksonville is going up 95 towards my hometown,” Dillon said. [SIGN UP: Action News Jax Daily Headlines Newsletter] Dillon says neither his case nor Richardson’s should have happened. “I don’t wish this upon my worst enemy. I was taken away from my home, taken away from my wife, taken way from my child. And it put in a holding cell for hours, put in the back of a van
The use of facial-scanning technology at several San Francisco gay bars has angered many in the LGBTQ+ community. At locations often frequented by people who are not publicly out, the notion of tracking patrons' whereabouts in the very havens sought for privacy raises many alarms. At least three gay bars in the Castro District have started using Patronscan Guard+, a technology intended to flag fake IDs, according to the San Francisco Gazetteer. But the experience of having a camera turned on every person entering the venues, often without warning, has left many outraged. Hart Owen told the outlet that she considers the surveillance technology a serious risk to privacy. “It’s really not great to have lists of gay people,” they said. Management at Mix, Badlands, and Toad Hall declined to speak to the Gazetteer about the issue, though bouncers pointed reporters to signage informing guests that the technology was being used. Related: Could Minneapolis welcome bathhouses back for the first time since the late 1980s? The Patronscan website offers insight into how the data can be used. The technology, according to the company, goes beyond simple document checks to validate identification and includes third-party checks. Those could flag people wanted by law enforcement or those appearing on sex offender registries. The company said it uses more than 8,500 forensic checks for every ID card scanned in bars, cross-referencing information to verify ages and names with listed addresses, “giving your front-of-house team real-time access to the patron intelligence they need.” The site also addresses privacy concerns. “We know this is one of the first questions operators ask, and we want to be clear: when a Third-Party Check runs, we use only the guest’s name and address to compare against publicly available information, such as confirming that an address exists and is
Spectrometer-free time-division multiplexed NIR time-of-flight vision system for visually similar material recognition - Open Access - 01.12.2026 - Article Abstract Introduction Machine vision systems have rapidly advanced across diverse fields, including automated manufacturing, robotics, and intelligent inspection. These developments have been largely driven by vision-based measurement (VBM) techniques, which integrate vision sensors, electronics, and computational algorithms1. The rise of artificial intelligence (AI), particularly machine learning (ML)-based classification and recognition methods, has significantly enhanced image analysis capabilities, improving the accuracy, reliability, and speed of measurement systems2‐7. Consequently, VBM has evolved from a simple image-processing approach into an essential tool for modern instrumentation and automation. Despite these advances, most existing vision systems still rely on RGB cameras, which provide only color and geometric information. This limits their ability to capture intrinsic material characteristics and fine surface geometry8‐11. RGB-D sensors partially alleviate this limitation by integrating active depth sensing, yet the mismatch between RGB and depth-map resolutions, along with insufficient spectral information, hinders their effectiveness in distinguishing visually similar materials12. Polarization-sensitive imaging has also been investigated as a useful modality for distinguishing visually similar materials by providing additional contrast related to scattering and surface optical properties13,14. Accurate and reliable acquisition of both material and geometric information remains a critical requirement for high-precision object recognition in production lines, recycling systems, and humanoid robotics15‐18. Anzeige The use of NIR lights (900–2500 nm) has emerged as a promising technique due to its rich material-dependent reflection and absorption features, which are more sensitive to molecular composition19‐23. Multispectral and hyperspectral imaging in the NIR range has demonstrated high effectiveness in identifying material characteristics, enabling more precise discrimination than visible-light imaging20,24‐27. However, these techniques typically rely on incoherent light sources, leading to low measurement efficiency and the inability to simultaneously capture depth information28‐31. NIR multispectral LiDAR systems have recently
NEWS Fort Myers man sues Jax Beach police, JSO after AI facial recognition leads to wrongful arrest As artificial intelligence becomes a bigger part of police investigations nationwide, critics warn that the technology can still make serious mistakes.
In today's newscast: Rally outside the Amazon data center in Gilroy tomorrow Protestors will rally outside the Amazon data center in Gilroy tomorrow afternoon against federal plans for an Immigration and Customs Enforcement facility nearby. Rebeca Armendariz previously helped organize the No Kings protest. She is connected to Santa Clara County's Community Agency for Resources, Advocacy and Services (CARAS) and the local Your Allied Rapid Response Network (YARR). "We want folks to remember the connection between ICE and data centers," said Armendariz, who adds Amazon provides cloud and data storage services to tech companies that work with ICE. These companies have provided facial recognition software and AI tools to identify potential targets for ICE arrests and deportations. For example, Armendariz says Amazon has partnered with tech company Palantir and more funding has emboldened the agency. "ICE now has an extremely large budget to hire and to build more infrastructure," said Armendariz. She also warns attempts to reopen a former federal women’s prison in Dublin, as Northern California’s first ICE detention center, could make it easier to carry out raids. "From the Central Coast to the Bay Area, having a processing center and having a detention center so nearby is gonna impact all of us. None of our communities is safe," she said. "So we need to push back and fight back everywhere from Dublin to Gilroy to the whole state and across the country.” California and Santa Clara County also fight back California Attorney General Rob Bonta and Santa Clara County have filed a joint lawsuit against the Trump administration over the proposed ICE facility in Gilroy. Santa Clara County counsel Tony LoPresti spoke at a press conference on Wednesday. "Our suit sends a simple message: in this county and in this state, the law doesn’t yield to the Trump
JACKSONVILLE, Fla. – As artificial intelligence becomes a bigger part of police investigations nationwide, critics warn that the technology can still make serious mistakes. News4JAX obtained arrest video showing a Fort Myers man being taken into custody after facial recognition software wrongly identified him as a child abduction suspect. Robert Dillon, 52, is now suing multiple law enforcement agencies in federal court, alleging his civil rights were violated after he was arrested and jailed before charges were later dropped. Body-worn camera video from August 2024 shows a Lee County Sheriff’s Office deputy at Dillon’s Fort Myers home telling him Jacksonville Beach police suspected him of luring or attempting to abduct a child. Police said Dillon was identified through facial recognition software, which they described as a match. Dillon was among more than two dozen people who have been wrongfully arrested after being identified through facial recognition technology since 2019, according to his attorneys. In Dillon’s case, the criminal charges were eventually dropped after further investigation determined police had arrested the wrong man. Dillon has filed a federal civil rights lawsuit against the Jacksonville Beach Police Department, the Jacksonville Sheriff’s Office — which his lawyers say ran the facial recognition search — and the Pinellas County Sheriff’s Office, which also houses AI technology. Police said the facial recognition system returned what they described as a 93% match. Ann Liebschutz, a government affairs specialist, said AI tools have improved rapidly in recent years. Still, she said mistaken arrests can be avoided when investigators apply proper standards before seeking an arrest warrant. “The professional duty of care that needs to be exercised when procuring an arrest warrant can eliminate any of these challenges we have with the technology identifying the wrong person,” Liebschutz said. She said she expects fewer mistaken arrests as agencies
Street Soldiers: Facial recognition In this week's episode of Street Soldiers, FOX 5 NY's Lisa Evers take a closer look at today's facials recognition technology with three experts: Jon Padfield, host of the Business Reform YouTube Channel, criminal defense attorney Phillip Hamilton and Ralph Cilento, co-founder of Facial Recognition Investigative Consultants.
Police will deploy live facial recognition at a multi-arts festival in Lowestoft. The technology will be used at the First Light Festival on Saturday, June 20, to identify individuals wanted in connection with serious offences. Suffolk police describe the move as part of broader efforts to embrace innovative tools that support investigations and prioritise public safety. Chief Inspector Andy Pursehouse said: "This innovative technology can play a vital role in protecting people, keeping them safe, investigating crimes and getting justice for victims. "It also means we are saving time and effort for our officers, leaving them able to do other work to protect and help the public. "The presence of the van is also a proven and powerful engagement tool." The forces says that last year more than 200 people came to speak to officers on the ground at the festival to understand more about the technology. Last year’s deployment also led to three arrests and engagement. Chief Inspector Pursehouse added: "I’m confident we will see a similar amount of interest this year as well. "Once again, we will be able to show the public how the system operates and show how the data is deleted and faces are blurred in less than a second if they are not matched against the watchlist. "It’s essential to say those attending the festival should feel reassured that we are only looking for those individuals who have committed serious offences." Live facial recognition compares faces captured on live camera feeds against a pre-determined ‘watch list’ of individuals of interest to police. Images not matching the list are deleted instantly and not stored but if a match is found, officers on the ground will verify the person’s identity and determine the appropriate next steps.
Man sues law enforcement alleging AI facial recognition technology led to wrongful arrest A Florida man is suing multiple law enforcement agencies after he alleges he was wrongfully arrested due to flawed AI facial recognition technology. "I don't wish this on my worst enemy," Robert Dillon, a father of one, told ABC News. Police body camera video footage, obtained by ABC News affiliate Gulf Coast News, shows Dillon's 2024 arrest outside his San Carlos Park home. "I was basically telling them this is crazy. I have no idea who did this, but it's not me," Dillon recalled of the arrest. Authorities had been searching for a suspect who allegedly tried to lure a child at a Jacksonville Beach restaurant, more than 300 miles away from Dillon's home. "To be accused of a heinous crime of that nature, I'm thinking I may not be coming back home," Dillon said. During their investigation, authorities allegedly fed "poor quality" surveillance images of the suspect into an AI-powered facial recognition program, which scanned facial features and found photos of the suspect and Dillon were allegedly "a 93% match," according to Dillon's lawsuit. Dillon said he knew he had been incorrectly identified, because he had never been to Jacksonville Beach. "I argued with the police officer for 20 minutes, and he insisted that 'I know that I'm looking at your mugshot,'" he said. The state attorney's office dropped Dillon's case weeks after his arrest, but Dillon said it took nearly a year to get the arrest wiped from his record, with help from the American Civil Liberties Union. Dillon and the ACLU are currently suing several law enforcement authorities and agencies, including the Pinellas County Sheriff's Office, which oversees the use of the AI facial recognition technology across Florida. "Despite [a] well-documented history of [facial
Erica Smith, CUSD Director of Communications and Community Relations, said that the camera installation schedule remains undecided, but said work will likely take place over summer to minimize the effect on campus activities. The cameras, made by tech company Verkada, use features like AI-powered facial recognition, license plate identification and text-based searching. Data is stored for 30 days, according to Verkada’s website. Data collected by the cameras belongs solely to the school district, unless the district gives explicit access to an outside party. The cameras are capable of sharing live footage with local law enforcement in emergency situations. The school board unanimously approved the camera replacements at a May 20 board meeting to only enable the camera’s vehicle history feature while disabling facial recognition features. However, the board didn’t discuss its ability to eventually enable the contentious facial recognition functions. Parents and students repeatedly expressed their concerns over privacy throughout CUSD board meetings, bringing student safety into question. Many cited Verkata’s numerous data leaks throughout the company’s history as a red flag. In 2021, the Federal Trade Commission sued the company for nearly $3 million after it failed to secure customer data, which led to a hacker accessing private customer cameras, according to the FTC website. Verkada is now annually audited by the FTC to ensure that security measures are continually updated and secure. REASONS FOR REPLACEMENT At the Feb. 18 board meeting, CUSD Director of Information Technology John Vincent spoke on why CUSD’s current cameras should be replaced — primarily, he said, because current systems are “antiquated” and approaching the end of its usability. After assessing different proposals, Vincent said using Verkada costs less than the current cameras used. He said that current cameras would cost $4.6 million to maintain and renew over a 10-year period, whereas Verkada will
1. Introduction Over the past decade, deep learning as a field has grown quite significantly, whether it be the compute capacity of hardware or the ingenuity behind architectures that utilize that hardware. But if you think about it for more than a second, the underlying architecture has remained consistent in a few key areas. We’ve seen a massive shift from convolutional networks to the new Transformer architectures that power today’s large language models, but the way these networks route information from one layer to another hasn’t changed all that much. Recently, researchers at DeepSeek-AI released a paper titled “mHC: Manifold-Constrained Hyper-Connections,” (Xie et al., 2025b)1 which proposes an entirely new redesign of this routing system. To really appreciate the solution they came up with, let’s look at how signal propagation has evolved over the past few generations of models, and why the current methods are hitting a wall. 2. The Backbone: Standard Residual Connections Firstly, to understand the specific problem that the authors are trying to solve, we need to talk about where it all started–The standard Residual Connection (He et al., 2015)2. Introduced back in 2015 with ResNets, the residual connection is arguably one of the most important architectural design choices used in every AI model out there. Mathematically, it looks like this: It simply means that the final output of a layer is the sum of its output and the input it originally got. The key component here is that bare xl term in the residual stream, which we call the identity mapping. It’s important because it acts as an uninterrupted pathway for the gradient signal to flow through the entire network from start to finish. This property is exactly what prevents gradients from vanishing or exploding during training and allows us to successfully train models with hundreds
Federal authorities are investigating large markings discovered in the grass on the National Mall that appear to spell out “86 47,” and the discovery quickly triggered a federal response that included evidence collection, testing and a review by multiple agencies. The markings were found Thursday near the World War II Memorial. Investigators have not determined how they were created, who made them or what message was intended. Download the SAN app today to stay up-to-date with Unbiased. Straight Facts™. Point phone camera here Investigators collect evidence Photographs taken from the Washington Monument by a Reuters photographer show what appears to be the sequence “8647” etched into the lawn. The eight is the easiest to identify, while the four and the seven are far less distinct, making the full message difficult to recognize from ground level. U.S. Park Police responded shortly before noon after receiving a vandalism report. Officers and investigators secured part of the area, collected grass samples and began examining what caused the discoloration. National Guard personnel were also seen at the site as authorities worked around several brown patches in the grass. The Interior Department characterized the incident as vandalism and said it would be investigated aggressively. “Any threat against the president is taken very seriously by the Department, and our U.S. Park Police will investigate this incident and hold those responsible accountable,” an Interior Department spokesperson said. Why the numbers matter The phrase “86 47” has become a flashpoint in political and legal disputes surrounding President Donald Trump. The number 47 refers to Trump as the nation’s 47th president. The meaning of “86” is more contested. Merriam-Webster traces the term to restaurant and service-industry slang commonly used to mean remove, reject or get rid of something. Trump allies and federal prosecutors have argued that, when paired with
U.S. Marines with the 2nd Marine Logistics Group color guard march during a change of command ceremony at Marine Corps Base Camp Lejeune, North Carolina, June 11, 2026. The change of command ceremony is a time-honored tradition which formally signifies the transfer of command and entails the total accountability, authority and responsibility from one individual to another. (U.S. Marine Corps photo by Sgt. Alfonso Livrieri) | Date Taken: | 06.11.2026 | | Date Posted: | 06.12.2026 05:11 | | Photo ID: | 9743325 | | VIRIN: | 260611-M-GD991-1093 | | Resolution: | 7593x5062 | | Size: | 6.46 MB | | Location: | MARINE CORPS BASE CAMP LEJEUNE, NORTH CAROLINA, US | | Web Views: | 24 | | Downloads: | 1 | This work, 2nd Marine Logistics Group Change of Command Ceremony [Image 12 of 12], by Sgt Alfonso Livrieri, identified by DVIDS, must comply with the restrictions shown on https://www.dvidshub.net/about/copyright.
Abstract Multilevel image thresholding is an important segmentation technique that partitions an image into meaningful regions in applications such as object recognition, medical imaging, and satellite image analysis. However, conventional techniques are limited by their poor sensitivity to initial conditions, lack of sufficient spatial contextual information, and slow convergence. The proposed method uses a novel hybrid model based on the Artificial Hummingbird Algorithm (AHA) to address the limitations of existing approaches. Here, Latin Hypercube, Sobol, Halton, and Sierpinski strategies are used during the population initialization phase to improve population diversity and search space coverage. This improves the exploration capability of the algorithm and supports better search space. The proposed methodology also uses spatial contextual information to improve the quality of segmentation. It also incorporates a relationship between neighboring pixels in order to retain more structure and enhance the visual performance of the output. For the exploitation phase, the Great Deluge Algorithm (GDA) is utilized as the optimization algorithm. Furthermore, the use of GDA serves as an adaptive acceptance function which reduces the chance of getting stuck during the search process. Minimum Cross Entropy Measure (MCEM) is used as the objective function to obtain optimal threshold values. Different evaluation metrics have been used to compare the results of the proposed method with other existing metaheuristic algorithms. The code is available at https://github.com/suprajatirumalasetti/AHA_GDA_Image_Segmentation_Code. Funding Open access funding provided by Vellore Institute of Technology- AP University. 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
Expert Q&A: AI & Food Safety Beyond the AI Hype AI is not a detection tool. It interprets patterns in data, which makes validation, governance, and human oversight essential to its use in food safety systems, says Willette Crawford, owner and principal of Katalyst Consulting LLC. This interview is part of our extended Q&A series exploring AI and food safety. As AI adoption grows, expectations for what these systems can do are not always aligned with reality. AI can identify patterns and highlight potential risk, but it cannot confirm contamination or replace human judgment. In this Q&A, Willette Crawford, owner and principal of Katalyst Consulting LLC, explains where the technology adds value and why validation and oversight are critical. Where do you see AI generating measurable value in food plants? AI is generating the most measurable value in areas where data is continuous, structured, and directly tied to decisions such as environmental monitoring, sanitation verification, and process control. The real impact isn’t automation, it’s prioritization. AI helps teams recognize meaningful signals earlier and act before a loss of control occurs. In that way, it strengthens preventive controls by improving timing and consistency, rather than replacing the system itself. Where does the hype around AI exceed the science? The hype exceeds the science when AI is positioned as a detection tool rather than what it actually is: a pattern recognition system built on imperfect data. AI doesn’t detect pathogens; it interprets signals that may correlate with risk. Without strong data quality, context, and validation, it can create a false sense of precision. So, the danger isn’t the technology itself—it’s overconfidence in what it can actually prove. The effectiveness of AI depends heavily on the quality of the data and the context in which the system is built. The real impact isn’t
A 93% police face match in Florida sent a Fort Myers crabber to jail after a child-luring case, but the arrest is fueling a bigger fight over one of the oldest facial-recognition systems in the US Synopsis Robert Dillon was wrongfully arrested after facial recognition software flagged him for a crime committed hundreds of miles away. This incident highlights the dangers of relying on AI for law enforcement, as a grainy photo led to an arrest warrant based solely on the technology's output. The ACLU lawsuit underscores the need for caution with these error-prone tools. How one grainy photo led to a wrongful arrest Police responded to reports of an attempted child abduction at a Jacksonville Beach McDonaldâs in November 2023. One witness said a man tried to persuade a girl under the age of 12 to go away with him. According to Reason magazine, the responding officer didnât even get a copy of the security footage; he just took cell phone pictures of the surveillance screen. Weeks passed with no leads. The investigating officer compared the photos to booking records and the sex offender registry but found no matches. Eventually, he sent the photos to other agencies for help, and thatâs when an investigator ran them through facial recognition software, which flagged Dillon. A police report reportedly said the software returned the match with high confidence. From there it was all downhill. According to the ACLU's official statement, Jacksonville Beach police based their arrest warrant solely on that facial recognition hit and a statement from a restaurant employee who picked Dillon's photo from a lineup. The issue? Dillon lived hundreds of miles away. Investigators even ran a license plate reader database, the suit says, which showed no sign of Dillonâs vehicles anywhere near Jacksonville Beach in the 48 hours