News

RisingAttacK quietly alters key features, tricking AI without changing the image’s appearanceVision systems in self-driving cars could be blinded by nearly invisible image modificationsThe attack fools top AI models used in cars, cameras, and healthcare diagnosticsArtificial intelligence is becoming more integrated into technologies that rely on visual recognition, from autonomous vehicles to medical imaging - but this increased utility also raises potential security risks, experts have warned.A new method called RisingAttacK could threaten the reliability of these systems by silently manipulating what AI sees.This could theoretically cause it to miss or misidentify objects, even when images appear unchanged to human observers.Targeted deception through minimal image alterationDeveloped by researchers at North Carolina State University, RisingAttacK is a form of adversarial attack that subtly alters visual input to deceive AI models.The technique does not require large or obvious image changes; instead, it targets specific features within an image that are essential for recognition.“This requires some computational power, but allows us to make very small, targeted changes to the key features that make the attack successful,” said Tianfu Wu, associate professor of electrical and computer engineering and co-corresponding author of the study.These carefully engineered changes are completely undetectable to human observers, making the manipulated images appear entirely normal to the naked eye.“The end result is that two images may look identical to human eyes, and we might clearly see a car in both images,” Wu explained.“But due to RisingAttacK, the AI would see a car in the first image but would not see a car in the second image.”This can compromise the safety of critical systems like those found in self-driving cars, which rely on vision models to detect traffic signs, pedestrians, and other vehicles.If AI is manipulated into not seeing a stop sign or another car, the consequences could be severe.The team tested the method against four widely used vision architectures: ResNet-50, DenseNet-121, ViTB, and DEiT-B. All four were successfully manipulated.“We can influence the AI's ability to see any of the top 20 or 30 targets it was trained to identify,” Wu said, citing common examples like cars, bicycles, pedestrians, and stop signs.While the current focus is on computer vision, the researchers are already looking at broader implications.“We are now in the process of determining how effective the technique is at attacking other AI systems, such as large language models,” Wu noted.The long-term aim, he added, is not simply to expose vulnerabilities but to guide the development of more secure systems.“Moving forward, the goal is to develop techniques that can successfully defend against such attacks.”As attackers continue to discover new methods to interfere with AI behavior, the need for stronger digital safeguards becomes more urgent.Via TechxploreYou might also likeDownloaded something dodgy? These are the best malware removal toolsNail the basics with the best firewalls available nowAnother top WordPress plugin hacked to allow account takeover

Apple is making another attempt to appeal the trade ban that forced it to remove the blood oxygen sensor from its smartwatches, Reuters reports. The company was forced to remove the feature in 2024, following a decision from the International Trade Commission (ITC) in 2023 that banned sales of the Apple Watch for violating health tech startup Masimo's patents.

The US Court of Appeals for the Federal Circuit heard new arguments from Apple and Masimo's lawyers on Monday over the legality of the original ITC ban. Apple's argument is that the fact Masimo was developing its own smartwatch when Apple released the Apple Watch Series 6 in 2020 — the first of the company's wearables with a blood oxygen sensor — shouldn't be enough to justify the ban. Masimo disagrees.

The appeals court has heard the company's arguments before, and even briefly paused the trade ban in December 2023 while the ITC considered Apple's request for a longer pause. Ultimately, the ITC ruled in Masimo's favor and denied Apple's request. Apple removed the blood oxygen sensor in 2024 so it could restart sales of the Apple Watch Series 9 and Apple Watch Ultra 2 in the US.

The companies' fight goes beyond just the blood oxygen sensor — Masimo's smartwatches were, ironically, found to infringe on Apple patents — but clearly Apple is hung up on the feature. The company's blood oxygen sensor was never as accurate as the pulse oximeters used in a doctor's office, but it was part of a fairly comprehensive suite of heart health features on the smartwatch. It's entirely possible Apple wants to reintroduce it in its watches as part of an even more ambitious health feature down the road.This article originally appeared on Engadget at https://www.engadget.com/wearables/apple-is-still-trying-to-overturn-the-ban-on-the-apple-watch-blood-oxygen-sensor-200135856.html?src=rss

Two flaws were first introduced in late 2013They reside in the Sudo command-line utilityPatches are available and users are advised to apply themTwo vulnerabilities were recently spotted in various Linux distributions which, when chained together, allow local attackers to escalate their privileges and thus run arbitrary files.The vulnerabilities are tracked as CVE-2025-32462 (severity score 2.8/10 - low severity), and CVE-2025-32463 (severity score 9.3/10 critical), and were found in the Sudo command-line utility for Linux and other Unix-like operating systems.All versions before 1.9.17p1 were said to be vulnerable, with Rich Mirch, the Stratascale researcher who found the flaws, saying they were lingering for more than a decade before being discovered. They were first introduced in late 2013, he added.Get 55% off Incogni's Data Removal service with code TECHRADAR Wipe your personal data off the internet with the Incogni data removal service. Stop identity thievesand protect your privacy from unwanted spam and scam calls.View DealA decade-old flawSudo (short for “superuser do”) is a command that allows a permitted user to execute a command as the root user or another user, as defined in the system’s security policy. It provides controlled administrative access without requiring users to log in as the root account.For example, a user might run a sudo command that installs Firefox on Ubuntu, since installing software system-wide usually requires administrative privileges."This primarily affects sites that use a common sudoers file that is distributed to multiple machines," Todd C. Miller, a maintainer for the Sudo project, said in an advisory. "Sites that use LDAP-based sudoers (including SSSD) are similarly impacted."The patch for Sudo was released in late June 2024, after responsible disclosure which happened in early April.Furthermore, different Linux distributions also released advisories, fixing the flaw for their variant of the OS. For CVE-2025-32462, these include AlmaLinux 8, AlmaLinux 9, Alpine Linux, Amazon Linux, Debian, Gentoo, Oracle Linux, Red Hat, SUSE, and Ubuntu, while for CVE-2025-32463, they include Alpine Linux, Amazon Linux, Debian, Gentoo, Red Hat, SUSE, and Ubuntu.Linux users are advised to apply the available patches and make sure their Linux desktop distributions are generally updated.Via The Hacker NewsYou might also likeOver 16 billion records leaked in "unimaginable" major data breach – here's what we knowTake a look at our guide to the best authenticator appWe've rounded up the best password managers