Hof – Today, artificial intelligence can describe images, recognize objects, and explain complex relationships. The pace of development is remarkable: so-called vision-language models (VLMs) combine text and image understanding in impressive ways. Yet, of all things, they struggle with a seemingly simple task—counting. Researchers at the Institute for Information Systems (iisys) at Hof University of Applied Sciences are now working to address this issue. “Many common models are very good at recognizing what can be seen in an image—but not reliably how many objects there are,” explains Prof. Dr. René Peinl from the Institute for Information Systems (iisys) at Hof University of Applied Sciences. Errors become more frequent when there are more than four or five objects of the same type. Why Counting Is So Difficult for AI The problem runs deeper than it may appear at first glance. While humans can intuitively grasp small quantities, larger numbers must be actively counted. This crucial step is missing in many AI models. In addition, existing training data is often unsuitable. “Some datasets are too simple and only encourage pattern recognition—others are too complex or flawed, for example due to occluded objects or ambiguous questions,” says institute director Prof. Peinl. As a result, models tend to “guess” or rely on learned expectations—sometimes producing surprisingly incorrect results. The Solution from Hof: An Artificial Dataset To tackle this problem in a targeted way, iisys has developed the SITUATE dataset. Instead of using real photographs, the researchers generate artificial 3D scenes with clearly defined properties. “We wanted to create an environment in which we can precisely control what happens in the image—and what does not,” says Prof. Dr. René Peinl. These scenes contain geometric objects such as cubes, spheres, or cylinders, with clearly defined positions (e.g., “to the left of the table”), allowing for targeted
May 4, 2026 · via idw-online.de
TheLostSwede News Editor - Joined - Nov 11, 2004 - Messages - 21,014 (2.68/day) - Location - Sweden | System Name | Overlord Mk MLI | |---|---| | Processor | AMD Ryzen 7 7800X3D | | Motherboard | Gigabyte X670E Aorus Master | | Cooling | Noctua NH-D15 SE with offsets | | Memory | 32GB Team T-Create Expert DDR5 6000 MHz @ CL30-34-34-68 | | Video Card(s) | Gainward GeForce RTX 4080 Phantom GS | | Storage | 1TB Solidigm P44 Pro, 2 TB Corsair MP600 Pro, 2TB Kingston KC3000 | | Display(s) | Acer XV272K LVbmiipruzx 4K@160Hz | | Case | Fractal Design Torrent Compact | | Audio Device(s) | SteelSeries Arctis Nova 3 Wireless | | Power Supply | be quiet! Pure Power 12 M 850 W | | Mouse | Logitech G502 Lightspeed | | Keyboard | Corsair K70 Max | | Software | Windows 11 Pro | | Benchmark Scores | https://valid.x86.fr/yfsd9w | Asustor Inc. today is announcing that it will unveil a range of new products at Computex 2026 in Taipei, to showcase its superior network storage prowress with numerous models in addition to its award winning lineup of network storage solutions. Flashstor Gen3 Series - Flagship Flash NAS for Creators with Optional AI The all-flash NAS, highly praised by professional photographers and video creators, is receiving a significant upgrade. The new Flashstor Gen3 series is equipped with an AMD Ryzen 5 Pro 8640U Six-Core processor, providing 16 TOPs of AI computing power. Enterprise-Grade Petabyte-Level Storage Environment - The Lockerstor R Pro Gen2 and the Xpanstor 12R Gen2 Join Forces To meet the stringent requirements of enterprises for high data storage and high availability, the newly launched Lockerstor 24R Pro Gen2 which is 4U and 24-bay in addition to the rest of the Lockerstor
May 4, 2026 · via techpowerup.com
Foremost, a passive data trail is information collected automatically by devices, networks, or systems as a byproduct of normal operation—like signals, logs, or metadata. Devices that leave passive data trails 1. Smartphones- These are the biggest contributors. As they offer- 1.) Cellular signals: Your phone constantly communicates with nearby cell towers (even when idle). 2.) Wi-Fi scanning: It probes for known networks, revealing device identifiers. 3.) Bluetooth: Emits signals that can be picked up by nearby devices (used in tracking beacons). 4.) Sensors: GPS, accelerometer, gyroscope generate location and movement data. Even with no apps open, your phone is still “talking” to networks. 2. Wearables (Smartwatches, Fitness Bands) Devices like smartwatches passively collect: > Heart rate and health data > Movement and sleep patterns > Location (if paired with phone or GPS-enabled) They sync periodically, creating background data logs. 3. Laptops & Computers a.) Connect to Wi-Fi networks and log IP addresses b.) Background services send diagnostic or usage data c.) Browsers track activity via cookies and scripts Even idle computers can generate network traffic. 4. Smart Home Devices (IoT) Examples: Smart speakers, Smart Televisions, Security or CCTV cameras, Smart thermostats They passively generate: 1.) Usage patterns (when you’re home, what you watch) 2.) Voice snippets (in some cases) 3.) Device interaction logs 5. Vehicles (Modern Connected Cars) Modern cars collect: A.) GPS location B.) Driving behavior (speed, braking) C.) System diagnostics Some connect to manufacturer servers or apps automatically. 6. Public Infrastructure Interactions Even without owning a device, passive trails happen via: i) CCTV cameras (facial recognition in some places) ii) Automatic toll systems (RFID tags) iii) Public Wi-Fi networks (device tracking via MAC addresses) 7. Payment Systems A.) Contactless cards and mobile wallets B.) Transaction logs (time, place, amount) Even a tap-to-pay leaves a digital trace. 8. Bluetooth
May 4, 2026 · via cybersecurity-insiders.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
May 4, 2026 · via youtube.com
AI and the Everyday: An Introduction to Focussing on Everyday Interactions of AI How can we highlight new ways in which researchers can further their engagement with AI in the Global South? During a workshop titled Public Debates, Everyday Injustice, and AI in the Majority World, a group of researchers and activists discussed how researchers should take power asymmetries, hierarchies, and inequalities into account, both across and within the global South. This introductory article kicks off the online continuation of their discussion. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Focussing on ‘the everyday’ gives us valuable insights into how AI shapes routine interactions in sites of work, leisure, education, and homes. It shows us the pervasiveness of AI-technologies and the various ways in which people negotiate it. Insights into the everyday provide an understanding of the messiness and the ambiguities that shape people’s responses to AI as they encounter its development, deployment, and use. In many of these interactions, engagements with AI are nested within discussions of opportunity, aspiration, and hope. These discussions are situated within a context where states and corporations propel a narrative of the inevitability of AI. However, people engage with AI from different standpoints. These standpoints include people's engagement with AI as actors with the agency to subvert, resist, and creatively challenge or use the power of AI, but also as those who are used, surveilled, and dominated through the ubiquitous use of AI-systems. What kinds of relations emerge with the use of AI? How can one capture and understand these relations? In this piece, we introduce and reflect on entries from an article symposium that arose from a workshop titled Public Debates, Everyday Injustice, and AI in the Majority World. These entries explore how people negotiate with AI in a
May 4, 2026 · via diggitmagazine.com
4 May 2026
UK report sparks dystopian fears with facial recognition
United Kingdom correspondent Diane To spoke to Melissa Chan-Green about how the UK government has said it might ban pro-Palestinian demonstrations as a move to tackle antisemitism and how a new report is sparking dystopian fears with facial recognition. She also spoke about King Charles' royal visit to Bermuda.
May 4, 2026 · via rnz.co.nz
You do not have permission to retrieve the URL or link you requested,
If you think this was a mistake please call 1300 134 174 or e-mail customercare@realestate.com.au
and quote the reference number #18.72cfdb17.1777882455.28ea31b3
May 4, 2026 · via realestate.com.au
In April, the Home Office announced £115m of new funding to create the “Police.AI” centre. Its aim is to introduce a “rapid and responsible” rollout of AI tools across all 43 forces in England and Wales. The National Police Chief Council (NPCC) claims it will establish UK policing as “a global leader in applied, responsible AI”. Currently, each police force in the UK largely determines for itself how and when to procure, trial and deploy AI tools, leading to a lack of standardisation and inefficient, costlier procurement practices. This has inevitably led to concerns being raised regarding policy, ethics and accountability – and nowhere has the debate been more pointed than around the use of Live Facial Recognition (LFR). This is because LFR is already being used at scale. More than 25,000 “retrospective” facial recognition (RFR) searches are carried out per month, with images from CCTV, mobile phones, dashcams, video doorbells and social media compared against a police national database of custody mugshots. In London, between January 2024 and September 2025, over 1,300 suspects were arrested with the help of live facial recognition technology (FRT). In February, the government concluded a consultation on legislation to govern the use of the technology by police. The Home Office signalled its intentions plainly: “The consultation will pave the way for new laws so all police forces can use this new technology with greater confidence and more often.” Up to this point, police advocates have stressed that the patchwork of checks and balances in place, such as the AI covenant in policing, and the Surveillance Camera Code of Practice, are sufficient for regulating the use of LFR. Critics counter that the public are being asked to rely on police assurances in what remains a legal grey area. Questions are also being raised about the
May 4, 2026 · via newstatesman.com
AI-powered image processing enhances daily life in Malaysia, from phone cameras to hospital scans and smart security systems. EVERY time you take a selfie and your phone magically enhances the lighting or when you walk past a security camera that seems to follow your movement, you are witnessing the quiet power of digital image-processing combined with artificial intelligence. These technologies are no longer just part of science fiction; they are part of our everyday lives, often working behind the scenes without us even noticing. Digital image-processing is a way for computers to “see” and “understand” pictures. It helps machines clean up blurry images, recognise faces and even detect objects. When paired with artificial intelligence, which allows computers to learn and make decisions, the results are impressive. Your phone knows where your face is in a photo. A hospital scanner can highlight signs of illness and a traffic camera can identify a speeding car in seconds. All of this is possible because of the way computers process images and learn from them. In Malaysia, these technologies are already being used in many sectors. Hospitals use AI to help doctors read X-rays more accurately. Security systems in shopping malls and offices use smart cameras to detect suspicious behaviour. Even agriculture is benefiting, with drones flying over fields to check crop health using image analysis. It is clear that AI and image processing are helping us live safer, healthier, and more efficient lives. But as these systems become smarter, they also raise important questions. Who controls the data collected by these cameras? How is your face being stored and used? Can these systems make mistakes? These are not just technical issues; they are ethical ones. For example, facial recognition can help catch criminals but it can also be used to track innocent people
May 4, 2026 · via thesun.my
Photos are an expected part of almost any visit to Disneyland. Whether it’s first-time international visitors or longtime returnees, just about everyone is clicking photos at some point. That everyone, however, is extending more and more to Disneyland personnel. My colleague Hannah Fry documented the growing use of facial recognition technology at the Happiest Place on Earth. Privacy experts are sounding the alarm, while some patrons are trying their best to avoid the technology. Let’s dive into her reporting to see what’s going on. Under the mouse’s eye Photographs of guests’ faces taken at the entrances to Disneyland and sister park California Adventure are run through biometric technology to convert the images into unique numerical values. The images can then be compared with pictures taken when a customer first used the ticket or annual pass. “Pretty much every other place is doing the same thing,” said John LeSchofs, 73, who visits the park roughly every six weeks with his wife. “The police, the government, they’re all using facial recognition. I don’t think it’s going to stop.” Disney officials say the technology helps make entering and reentering the park easier and prevents fraud. Sounding the alarm The rapid growth of facial recognition over the last decade has raised concerns among privacy experts who caution that such data can easily be turned over to law enforcement entities or make companies hacking targets. “The normalization of facial surveillance is really problematic,” said Ari Waldman, a professor of law at UC Irvine. “We can’t go around life hiding our faces, so this isn’t just [the] next step in surveillance; it’s qualitatively different. In a world of facial recognition, when people leave their house, it automatically means they’re identified.” Following a familiar pattern Venues over the last decade have increasingly relied on facial recognition to
May 4, 2026 · via union-bulletin.com
Why Facial Recognition Technology Demands A Warrant Facial recognition technology (FRT) seems to be everywhere these days—it is in malls, airports, and our ever-present smartphones. First commercialized in the mid-1990s, FRT today is used widely in the police investigation process due to its efficiency. Despite how commonly the technology is used as an investigative tool, the legal protections to govern its use and avoid government overreach have not been updated accordingly due to a lack of federal FRT legislation. The Eighth District Court of Appeals case State v. Tolbert (2025) reveals this lack of judicial oversight over FRT. While the case is still active and ongoing, the facts established in the case and the rulings to date point to the need for clearer FRT regulations. In State v. Tolbert, the Cleveland Police Department used a facial recognition match from a fusion center to identify Qeyeon Tolbert as a suspect for a homicide case. From this identification match, Homicide Detective Michael Legg secured a search warrant of an apartment linked to Tolbert, resulting in Tolbert’s arrest. Although the Eighth District Court of Appeals found that the use of FRT to obtain a warrant did not violate Tolbert’s Fourth Amendment right to a truthful affidavit warrant, the court’s ruling overlooks an important concern that the defendant raises: the unregulated use of FRT to identify a suspect. Given that FRT is an efficient surveillance tool becoming more commonly used in criminal investigation, lack of government regulation can encourage unreasonable searches in violation of the Fourth Amendment. To protect citizens from unregulated mass surveillance, while also acknowledging the benefits of FRT, the use of FRT to verify or identify an individual should be treated as a search under the Fourth Amendment. State v. Tolbert focuses on defendant Qeyeon Tolbert, a twenty-three-year-old African American man
May 3, 2026 · via culawreview.org
Facial recognition apps that allow police officers to carry out on-the-spot identity checks with their phones are being quietly adopted by UK forces before a potential national rollout. At least one force has also discussed using facial recognition technology with live-streamed body-worn video footage and using drone-mounted cameras to identify people at protests. The developments mark a significant expansion of the police’s use of facial recognition which, has to date, largely relied on cameras in public spaces and searches using images taken from sources such as CCTV and social media. Merseyside police is one of several forces to give officers access to handheld facial recognition technology, known as operator-initiated facial recognition (OIFR). OIFR allows officers to take photos of faces and search for matches on police databases. It is typically used when someone refuses to provide their identity or officers believe they have given a false name, although forces say it can also be used to check the identities of those who are unconscious or otherwise unable to provide details. Operator-initiated facial recognition is one of three types available to police. Retrospective facial recognition is the oldest form and is used to search images from crime scenes against pictures taken of people on arrest. Polling has found it to be the most acceptable form of facial recognition among the public. Live facial recognition uses live video footage of people passing cameras and comparing their faces to wanted lists. In December the government announced an expansion of facial recognition technology with the aim for it to be used daily by all 43 police forces in England and Wales. More than a dozen already use it routinely. Sarah Jones, the policing minister, wants it adopted by all of the forces, likening the impact on crime to the breakthrough of DNA matching in
May 3, 2026 · via thetimes.com
Photos are an expected part of almost any visit to Disneyland. Whether it’s first-time international visitors or longtime returnees, just about everyone is clicking photos at some point. That everyone, however, is extending more and more to Disneyland personnel. My colleague Hannah Fry documented the growing use of facial recognition technology at the Happiest Place on Earth. Privacy experts are sounding the alarm, while some patrons are trying their best to avoid the technology. Let’s dive into her reporting to see what’s going on. Under the mouse’s eye Photographs of guests’ faces taken at the entrances to Disneyland and sister park California Adventure are run through biometric technology to convert the images into unique numerical values. The images can then be compared with pictures taken when a customer first used the ticket or annual pass. “Pretty much every other place is doing the same thing,” said John LeSchofs, 73, who visits the park roughly every six weeks with his wife. “The police, the government, they’re all using facial recognition. I don’t think it’s going to stop.” Disney officials say the technology helps make entering and reentering the park easier and prevents fraud. Sounding the alarm The rapid growth of facial recognition over the last decade has raised concerns among privacy experts who caution that such data can easily be turned over to law enforcement entities or make companies hacking targets. “The normalization of facial surveillance is really problematic,” said Ari Waldman, a professor of law at UC Irvine. “We can’t go around life hiding our faces, so this isn’t just [the] next step in surveillance; it’s qualitatively different. In a world of facial recognition, when people leave their house, it automatically means they’re identified.” Following a familiar pattern Venues over the last decade have increasingly relied on facial recognition to
May 3, 2026 · via thederrick.com
Artificial intelligence (AI) road safety cameras have been rolling out across Australia, resulting in a large number of fines. For example, roughly 184,000 infringements have been issued in Western Australia since the cameras were launched in October last year. In New South Wales, more than 130,000 fines were issued in 2024-2025, the first year the technology was used. In Queensland, about 114,000 fines based on AI image recognition technology were issued in 2024. But it is the size of the penalties that has drawn much attention and criticism. In WA, penalties start at A$550 and four demerit points for seatbelt infringements. These can add up quickly in cases where people get repeated fines. For example, one WA driver with a seatbelt exemption was issued almost A$20,000 in fines after multiple seatbelt infringements. Although fines can be challenged, there are growing concerns about the accessibility and fairness of the appeal process, especially at scale. The technology also raises the question of whether AI camera systems are the best way to promote safe driving. How do AI road safety cameras work? AI road safety cameras use computer vision systems to identify patterns in still images captured by roadside cameras. AI software initially reviews all images. Where no infringement is detected, the image is automatically deleted. Where potential infringements are detected, the images are then flagged for human review before an infringement is sent out to a driver. The approach is intended to increase efficiency, as it saves officers from having to manually review large volumes of images. Drivers can challenge these fines where they believe an error occurred, or in cases where a valid excuse can be established. In Queensland, a driver successfully challenged a fine he received after his passenger shifted their seatbelt under their arm mid-trip. Representing himself in court,
May 3, 2026 · via theconversation.com
Photos are an expected part of almost any visit to Disneyland. Whether it’s first-time international visitors or longtime returnees, just about everyone is clicking photos at some point. That everyone, however, is extending more and more to Disneyland personnel. My colleague Hannah Fry documented the growing use of facial recognition technology at the Happiest Place on Earth. Privacy experts are sounding the alarm, while some patrons are trying their best to avoid the technology. Let’s dive into her reporting to see what’s going on. Under the mouse’s eye Photographs of guests’ faces taken at the entrances to Disneyland and sister park California Adventure are run through biometric technology to convert the images into unique numerical values. The images can then be compared with pictures taken when a customer first used the ticket or annual pass. “Pretty much every other place is doing the same thing,” said John LeSchofs, 73, who visits the park roughly every six weeks with his wife. “The police, the government, they’re all using facial recognition. I don’t think it’s going to stop.” Disney officials say the technology helps make entering and reentering the park easier and prevents fraud. Sounding the alarm The rapid growth of facial recognition over the last decade has raised concerns among privacy experts who caution that such data can easily be turned over to law enforcement entities or make companies hacking targets. “The normalization of facial surveillance is really problematic,” said Ari Waldman, a professor of law at UC Irvine. “We can’t go around life hiding our faces, so this isn’t just [the] next step in surveillance; it’s qualitatively different. In a world of facial recognition, when people leave their house, it automatically means they’re identified.” Following a familiar pattern Venues over the last decade have increasingly relied on facial recognition to
May 3, 2026 · via dailygazette.com
Photos are an expected part of almost any visit to Disneyland. Whether it’s first-time international visitors or longtime returnees, just about everyone is clicking photos at some point. kAm%92E 6G6CJ@?6[ 9@H6G6C[ :D 6IE6?5:?8 >@C6 2?5 >@C6 E@ s:D?6J=2?5 A6CD@??6=]k^Am kAm|J 4@==628F6 k2 9C67lQ9EEADi^^HHH]=2E:>6D]4@>^A6@A=6^92??29\7CJQmw2??29 uCJk^2m 5@4F>6?E65 E96 k2 9C67lQ9EEADi^^HHH]=2E:>6D]4@>^42=:7@C?:2^DE@CJ^a_ae\_c\ag^5:D?6J=2?5\C@==D\@FE\724:2=\C64@8?:E:@?\2E\A2C<\6?EC2?46D\96C6D\9@H\:E\H@C<DQm8C@H:?8 FD6 @7 724:2= C64@8?:E:@? E649?@=@8Jk^2m 2E E96 w2AA:6DE !=246 @? t2CE9]k^Am kAm!C:G24J 6IA6CED 2C6 D@F?5:?8 E96 2=2C>[ H9:=6 D@>6 A2EC@?D 2C6 ECJ:?8 E96:C 36DE E@ 2G@:5 E96 E649?@=@8J]k^Am kAm{6E’D 5:G6 :?E@ 96C C6A@CE:?8 E@ D66 H92E’D 8@:?8 @?]k^Am kAm&?56C E96 >@FD6’D 6J6k^Am kAm!9@E@8C2A9D @7 8F6DED’ 7246D E2<6? 2E E96 6?EC2?46D E@ s:D?6J=2?5 2?5 D:DE6C A2C< r2=:7@C?:2 p5G6?EFC6 2C6 CF? E9C@F89 3:@>6EC:4 E649?@=@8J E@ 4@?G6CE E96 :>286D :?E@ F?:BF6 ?F>6C:42= G2=F6D]k^Am kAm%96 :>286D 42? E96? 36 4@>A2C65 H:E9 A:4EFC6D E2<6? H96? 2 4FDE@>6C 7:CDE FD65 E96 E:4<6E @C 2??F2= A2DD]k^Am kAm“!C6EEJ >F49 6G6CJ @E96C A=246 :D 5@:?8 E96 D2>6 E9:?8[” D2:5 y@9? {6$49@7D[ fb[ H9@ G:D:ED E96 A2C< C@F89=J 6G6CJ D:I H66<D H:E9 9:D H:76] “%96 A@=:46[ E96 8@G6C?>6?E[ E96J’C6 2== FD:?8 724:2= C64@8?:E:@?] x 5@?’E E9:?< :E’D 8@:?8 E@ DE@A]”k^Am kAms:D?6J @77:4:2=D D2J E96 E649?@=@8J 96=AD >2<6 6?E6C:?8 2?5 C66?E6C:?8 E96 A2C< 62D:6C 2?5 AC6G6?ED 7C2F5]k^Am kAm$@F?5:?8 E96 2=2C>k^Am kAm%96 C2A:5 8C@HE9 @7 724:2= C64@8?:E:@? @G6C E96 =2DE 564256 92D C2:D65 4@?46C?D 2>@?8 AC:G24J 6IA6CED H9@ 42FE:@? E92E DF49 k2 9C67lQ9EEADi^^HHH]=2E:>6D]4@>^42=:7@C?:2^DE@CJ^a_ae\_`\_e^?6H\42=:7@C?:2\E@@=\42?\DE@A\4@==64E:@?\D2=6\@7\@?=:?6\A6CD@?2=\52E2Qm52E2 42? 62D:=J 36 EFC?65 @G6Ck^2m E@ =2H 6?7@C46>6?E 6?E:E:6D @C >2<6 4@>A2?:6D 924<:?8 E2C86ED]k^Am kAm“%96 ?@C>2=:K2E:@? @7 724:2= DFCG6:==2?46 :D C62==J AC@3=6>2E:4[” D2:5 pC: (2=5>2?[ 2 AC@76DD@C @7 =2H 2E &r xCG:?6] “(6 42?’E 8@ 2C@F?5 =:76 9:5:?8 @FC 7246D[ D@ E9:D :D?’E ;FDE ,E96. ?6IE DE6A :? DFCG6:==2?46j :E’D BF2=:E2E:G6=J 5:776C6?E] x? 2 H@C=5 @7 724:2= C64@8?:E:@?[ H96? A6@A=6 =62G6 E96:C 9@FD6[ :E 2FE@>2E:42==J >62?D E96J’C6 :56?E:7:65]”k^Am kAmu@==@H:?8 2 72>:=:2C A2EE6C?k^Am kAm'6?F6D @G6C E96 =2DE 564256 92G6 :?4C62D:?8=J C6=:65 @?
May 3, 2026 · via swoknews.com
China Smart Vision Processing Chips Market 2026 Analysis and Forecast to 2035 Executive Summary Key Findings - The China Smart Vision Processing Chips market is projected to grow from approximately USD 4.5–5.5 billion in 2026 to over USD 14–18 billion by 2035, representing a compound annual growth rate (CAGR) of roughly 13–16% driven by the country's dominant position in surveillance, automotive, and consumer electronics manufacturing. - China's reliance on imported advanced-node chips (7nm and below) remains high, with over 60–70% of high-performance Smart Vision Processing Chips sourced from Taiwan, South Korea, and the United States, creating a structural supply vulnerability amid ongoing export control restrictions. - Domestic fabless design houses and integrated device manufacturers (IDMs) have captured an estimated 30–40% of the Chinese market by volume in 2025–2026, primarily in mid-range and edge-deployed vision SoCs, but remain constrained by limited access to leading-edge foundry capacity and advanced packaging substrates. Market Trends Observed Bottlenecks Access to advanced semiconductor foundry capacity Licensing of critical AI/vision IP blocks Long OEM qualification cycles (especially automotive) Shortage of specialized chip design engineers Supply of advanced packaging substrates - Edge AI inference is rapidly displacing cloud-centric vision processing: by 2026, over 55–65% of all Smart Vision Processing Chips shipped in China are expected to perform real-time inference on-device, driven by latency requirements in autonomous driving, industrial inspection, and smart city surveillance. - Automotive ADAS and in-cabin monitoring has become the fastest-growing application segment, with China's vehicle production exceeding 28 million units annually and regulatory mandates for driver monitoring systems (DMS) accelerating adoption of dedicated vision processors. - Integration of Convolutional Neural Network (CNN) accelerators and Tensor core engines directly onto vision SoCs has become standard, with chip prices for mid-range edge AI vision processors declining by 8–12% per year as competition intensifies among domestic and international
May 3, 2026 · via indexbox.io
Google announced on April 29, 2026, a new feature for Google Photos that uses artificial intelligence to automatically scan a user's photo library, extract the clothing items that appear across those images, and assemble them into a dedicated digital wardrobe. The feature is scheduled to begin rolling out during the summer of 2026, first on Android and then on iOS, according to Tommy Meaney, Senior Product Manager at Google Photos. The announcement arrives as Google has spent nearly two years expanding AI-powered visual tools across its product line - from Google Lens shopping integrations to virtual try-on inside AI Mode in Search and, most recently, Circle to Search gaining the ability to identify every item in a complete outfit at once. The Google Photos wardrobe feature is distinct from all of those because it operates entirely on clothing the user already owns, not on items for sale elsewhere. How the AI wardrobe cataloging works According to the announcement, the wardrobe feature constructs its catalog by analyzing the photos already stored in a user's Google Photos library. The AI identifies clothing items visible in those images and classifies them by category. The categories listed in the announcement include jewelry, tops, and bottoms - suggesting the classification system covers a broad range of garment types rather than a narrow subset. Once the library has been processed, the resulting collection becomes browsable by category. A user can filter to see all tops together, all bottoms together, or all jewelry, rather than only having access to images organized by date or location as the current interface provides. The intended effect, according to the announcement, is that items buried deep in a photo archive become discoverable again, even if they appear in a photo taken years earlier. The depth of that library scan is notable.
May 3, 2026 · via ppc.land
Britain’s biometrics watchdogs have warned that national oversight of AI-powered face scanning to catch criminals is lagging far behind the technology’s rapid growth. With the Metropolitan police almost doubling the number of faces they scan in London over the past 12 months and a rising use of the technology by retailers in the UK, Prof William Webster, the biometrics commissioner for England and Wales, said the “slow pace of legislation was trying to catch up with the real world” and “the horse had gone before the cart”. Dr Brian Plastow, who holds the same role in Scotland, warned the technology was “nowhere near as effective as the police claim it is” and said there was a “patchwork legal framework” throughout the UK. He said in England and Wales, police were “really just marking their own homework”. The watchdogs said new laws were needed to govern when and how police forces used live facial recognition technology, with a new regulator to clamp down on misuse. Several bodies have oversight of the technology, including the Information Commissioner’s Office (ICO) and the Equality and Human Rights Commission. The Home Office is considering a new legal framework for the technology as it also plans to introduce nationally what it calls “the biggest breakthrough for catching criminals since DNA matching”. Members of the public wrongly labelled as suspected criminals by shops using AI cameras said there was no accountability or recourse to complain. They said the system had left them feeling “guilty until proven innocent”. They described the ICO, which is responsible for monitoring facial recognition tech and the biometric data it uses, as “toothless” and unresponsive. British police forces and high street retailers claim the technology makes streets safer, but others criticise it as Big Brother-style mass surveillance, with risks for civil liberties and
May 3, 2026 · via theguardian.com
When Ian Clayton, a retired health and safety professional from Chester, popped into Home Bargains one February lunchtime, he was suddenly approached by a stern-looking member of staff. “Excuse me, can you please put everything down and leave the shop now?” she said. Clayton recalled how he was stunned, and it was only as he was briskly walked past the tills towards the exit that he stopped to ask what he had done. “You’ve come up on our system called Facewatch as a shoplifter,” came the reply. “There’s a poster in the window.” With that, he was left outside the shop alone, with a QR code to scan and no idea what had happened. He is one of a number of people who have spoken to the Guardian after being falsely identified as a thief by shops using Facewatch, a live facial recognition system being rolled out across the UK to clamp down on retail crime. The company’s website claims that its system has a 99.98% accuracy rate and that last month it sent 50,288 alerts of “known offenders” to shops including B&M, Home Bargains, Sports Direct, Farm Foods and Spar, which all now use the software. But those who have been wrongly identified and forced to leave shops, either via the technology itself or human error, say they were given no support, and did not know how to complain about their treatment or prove their innocence. Clayton, 67, said that after he was ejected from Home Bargains he tried calling a phone number on a Facewatch poster, and was sent through to a message saying the company did not take calls and he had to send an email instead. He was only able to get answers after submitting a subject access request – a formal request under data protection
May 3, 2026 · via theguardian.com