Tuesday, November 17, 2026

Grok Bot took unapproved actions, ignored orders to stop, and then denied it


<p>Researchers recently discovered unexpected behavior from Grok Bot, the AI agent platform from Elon Musk’s SpaceXSI. Over five days, Grok bots installed on two separate employee laptops made 31 requests to access data stored in employee calendar, email, Slack, and GitHub accounts. The bots performed the actions even though the employees had never ordered them and continued doing them even after employees instructed them to stop. The bots also denied having taken the actions at all.</p> <p>The behavior was <a href="https://newcore.com/grok-bot-report">reported Tuesday</a> by NewCore, the maker of a managed gateway product that acts as an intermediary between AI agents and the tools and accounts they are assigned to work with. In some cases, the unapproved access attempts occurred while the laptop with Grok Bot installed had its lid shut, or late at night while the employee was sleeping. None of the unprompted actions were recorded in Grok logs. Without the gateway logs, the employees would have had no idea they occurred. And of course, without the gateway, the unapproved accesses would have succeeded—again without the end user's knowledge.</p> <h2>It’s “worse than a bug”</h2> <p>Once the employees learned that Grok Bot had tried to check in with accounts without prompting, one of them instructed the agent, four separate times, to stop taking any actions on its own. The agent's reply was: “I won’t set that up, and I’ll stop offering recurring check-ins.” Despite acknowledging the explicit instructions, the Grok Bot made more unapproved access attempts. When one of the employees asked the bot if it made any accesses, it checked twice and then outputted: “Overnight into Friday morning I was idle.”</p><p><a href="https://arstechnica.com/security/2026/11/grok-bot-took-unapproved-actions-ignored-orders-to-stop-and-then-denied-it/">Read full article</a></p> <p><a href="https://arstechnica.com/security/2026/11/grok-bot-took-unapproved-actions-ignored-orders-to-stop-and-then-denied-it/#comments">Comments</a></p> Reference : https://ift.tt/on0f3MG

Understand How Satellite Links Shape IoT Device Design and Testing


<img src="https://spectrum.ieee.org/media-library/gray-anritsu-logo-with-the-tagline-advancing-beyond-on-transparent-background.png?id=68666627&width=980"/><br/><br/><ul><li>How 3GPP Releases 15, 16, and 17 brought satellite access into 5G NR, NB-IoT, and eMTC.</li><li>Why satellite links suit IoT devices that send small, infrequent, and delay-tolerant messages.</li><li>Where non-terrestrial networks (NTN) add value, including asset tracking, agriculture, public safety, and offshore operations.</li><li>How power efficiency, RF signaling, and antenna design shape satellite-connected device development.</li><li>Why long propagation delay, Doppler shift, and intermittent coverage make static test methods insufficient.</li><li>Which measurements show deployment readiness, such as transmit power, modulation accuracy, spectral emissions, receiver sensitivity, and block error rate (BLER).</li><li>How emulated satellite conditions help verify synchronization, signaling, HARQ behavior, mobility, and interoperability.</li></ul><div><span><a href="https://content.knowledgehub.wiley.com/5g-nr-ntn-design-and-test-challenges-for-satellite-connected-iot/" target="_blank">Download this free whitepaper now!</a></span></div> Reference: https://ift.tt/7KvRgEl

Monday, November 16, 2026

Shareholders sue Intel, its CEO and CFO after stock drops 28.6 percent in one day


<p>Intel is being sued by shareholders, who are accusing the chipmaker of misleading them about the impact of the <a href="https://arstechnica.com/information-technology/2026/10/pc-shipments-fall-20-1-percent-in-sharpest-decline-since-q1-2023/">PC sales slump </a>that contributed to a drop of 28.6 percent in the price of Intel's stock on October 30.</p> <p>The lawsuit (<a href="https://cdn.arstechnica.net/wp-content/uploads/2026/11/Intel-complaint.pdf">PDF</a>), filed on Friday in the San Jose, California, federal court, is seeking class-action certification and lists the City of Southfield Fire and Police Retirement System and “all others similarly situated” as plaintiffs. In addition to Intel, Intel CEO Lip-Bu Tan and Intel’s EVP and CFO David Zinsner are listed as defendants.</p> <p>The plaintiffs claim that Intel made “materially false and misleading statements and omissions” between July 23 and October 29. They are seeking unspecified compensatory damages, plus legal and other associated fees, for shareholders during that period.</p><p><a href="https://arstechnica.com/tech-policy/2026/11/shareholders-sue-intel-its-ceo-and-cfo-after-stock-drops-28-6-percent-in-one-day/">Read full article</a></p> <p><a href="https://arstechnica.com/tech-policy/2026/11/shareholders-sue-intel-its-ceo-and-cfo-after-stock-drops-28-6-percent-in-one-day/#comments">Comments</a></p> Reference : https://ift.tt/42V8uSp

Advanced Metrology for Defect Detection, Tool Health, and Semiconductor Yield Enhancement


<img src="https://spectrum.ieee.org/media-library/evident-logo-with-stylized-triple-line-v-on-a-transparent-checkerboard-background.png?id=68617442&width=980"/><br/><br/><p>In advanced semiconductor manufacturing, small defects can have a large effect on yield. Thin film deposition tools must stay stable and qualified. Process drift must be caught early. When a defect does appear, engineers need to know where it came from and how it affects device performance. This webinar shows how in-line inspection and offline characterization combine into one workflow, from tool monitoring to root-cause analysis.</p><p>In the first half, Dr. Bitan Chakraborty from Intel, explains how surface scan metrology fits into station monitors for thin film tools. The session shows how baseline signatures and real-time monitoring reveal hardware wear, particle generation, and process drift before they reach production wafers. It then moves to in-line scans on patterned wafers. These scans expose layer-level defects and show how they form and spread through multi-layer device stacks. Linking this defect data to electrical test results helps teams find root causes and focus their yield improvement work.</p><p>In the second half, Dr. Jacob Petersen from Evident Scientific, addresses a common question: what happens after an in-line tool flags a problem? In-line tools are fast, but they cannot always provide the details needed for a firm conclusion. Offline 3D optical profilometry fills that gap by combining three complementary inspection techniques in a single system. It delivers dimensional measurements and helps determine whether a defect lies within the sample or comes from foreign material. The session covers how several optical techniques support inspection from bare silicon to fully packaged chips. It also shows how automation enables fast measurement of features down to the sub-nanometer scale with minimal operator handling.</p><p><strong>What You Will Learn:</strong></p><ul><li>How surface scan metrology in station monitors tracks tool health and supports the qualification of thin film deposition processes</li><li>How baseline signatures and real-time monitoring detect particle generation, hardware degradation, and process drift before production is affected</li><li>How in-line scans on patterned wafers reveal layer-level defects and show how they propagate through multi-layer device structures</li><li>When offline 3D optical profilometry should follow in-line inspection, and how it measures dimensions and separates sample defects from foreign material</li><li>How correlating tool health data, inspection results, and electrical performance links physical defects to yield loss and guides targeted improvement projects</li></ul><div><span><a href="https://event.on24.com/wcc/mr/7058967/655F0A675F5C3415546D597247025DA8?utm_source=evident122&utm_medium=IEEE&utm_campaign=series1" target="_blank">Register now for this free webinar!</a></span></div> Reference: https://ift.tt/6iLsVRh

28 Zero-Day Findings in Under Two Days. Is Robotics Security Keeping Pace?


<img src="https://spectrum.ieee.org/media-library/humanoid-and-dog-robots-watch-a-massive-swirling-storm-over-a-modern-city-walkway.jpg?id=68562173&width=1200&height=800&coordinates=38%2C0%2C38%2C0"/><br/><br/><p><em>This article is brought to you by <a href="https://vicone.com/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-10" rel="noopener noreferrer" target="_blank">VicOne</a>.</em></p><p>VicOne identified an initial set of <a href="https://www.zerodayinitiative.com/advisories/upcoming/" target="_blank">28 zero-day findings</a> in less than two days using AI-assisted research across software and tools supporting the Physical AI ecosystem. These include projects developed or maintained by Google, NVIDIA, and Meta, among others.</p><p>Based on initial validation, all 28 findings are classified as zero-day vulnerabilities. The count may change following vendor and vulnerability-authority review, and some findings do not yet have CVE tracking numbers.</p><p>Many of these tools help teams train, simulate, test, and operate robots. Weaknesses in them could influence the models teams build, the test results they trust, and the instructions sent to robots.</p><p>These findings create two demands for robotics teams: responding as <a href="https://vicone.com/blog/how-mythos-is-reshaping-vulnerability-management-what-cisos-need-to-know-about-vulnops/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" rel="noopener noreferrer" target="_blank"><span>AI helps uncover weaknesses faster</span></a>, and <a href="https://vicone.com/blog/can-a-robot-do-everything-right-and-still-be-unsafe/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" target="_blank"><span>assessing how each one could affect robot safety and operations</span></a>.</p><h2>What AI changes about vulnerability discovery</h2><p>Vulnerability discovery involves examining code, tracing how systems communicate, and testing whether a suspected weakness can actually be exploited. AI can accelerate parts of this work, helping researchers analyze more software in less time. Attackers could use similar capabilities.</p><p>When the weakness is a <strong>zero-day </strong>— previously unknown and without an available fix — faster discovery can leave teams with little time to respond. While that exposure does not necessarily mean an attack has occurred, teams still need to determine whether they use the affected software and what an attacker could access or change.</p><h2>Where the findings sit across the robot lifecycle</h2><p>VicOne’s findings span three stages of the robotics lifecycle: development and training, simulation and validation, and deployment and runtime. Weaknesses at each stage could affect what a robot learns, how it is tested, and how it is controlled.</p><p class="shortcode-media shortcode-media-rebelmouse-image"> <img alt="Bar chart: 50% development, 25% simulation, 25% deployment in robotics workflow." class="rm-shortcode" data-rm-shortcode-id="3798c8e581f193e51391a754e455a9d8" data-rm-shortcode-name="rebelmouse-image" id="f075f" loading="lazy" src="https://spectrum.ieee.org/media-library/bar-chart-50-development-25-simulation-25-deployment-in-robotics-workflow.jpg?id=68562185&width=980"/> <small class="image-media media-caption" placeholder="Add Photo Caption...">Figure 1. In this research set, 75% of findings involved development, training, simulation, or validation tools, highlighting risks that can arise before a robot is deployed.</small><small class="image-media media-photo-credit" placeholder="Add Photo Credit...">VicOne</small></p><p>The following scenarios show how a compromised component could influence the robot at each stage.</p><ul><li><strong>Development and training: Hidden changes can reach the finished robot<br/></strong>A team downloads a pretrained model to adapt for its robot. Loading a malicious file could give an attacker access to the training system, allowing changes to training data or the model itself. A manipulated model could behave normally during routine testing but respond incorrectly to a particular object, scene, or command after deployment. The risk can enter through a tool the team trusts, before anyone attacks the robot directly.</li><li><strong>Simulation and validation: A passing test could give false confidence<br/></strong>Importing a compromised scene or robot model could give an attacker access to the simulation environment. With sufficient control, the attacker could alter simulated sensor readings, physical conditions, or test results. A robot could then appear to pass tests that should have exposed unsafe behavior, leading the team to approve it for real-world use on the basis of misleading evidence.</li><li><strong>Deployment and runtime: Operators could lose visibility or control<br/></strong>An attacker could exploit weaknesses to change files that influence motion planning or send malicious map data that crashes the operator’s monitoring display. Compromising a remote-operation computer could also put the instructions sent to the robot at risk. Depending on the system and its safeguards, these weaknesses could alter robot movement or leave operators unable to see what is happening and respond promptly.</li></ul><p>These are potential consequences and not demonstrated outcomes for every finding. Actual effects depend on the attacker’s access, the system’s design, and whether other safeguards prevent the changes from reaching physical behavior.</p><p>In a separate demonstration, <a href="https://www.linkedin.com/showcase/vicone-lab-r7/posts/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" target="_blank"><span>VicOne researchers</span></a> provide a concrete example of a weakness affecting a protective function. The researchers exploited a vulnerability to prevent a robot’s remote-stop command from working. The case shows why teams need to test whether safeguards remain effective when the software supporting them is compromised.</p><p class="shortcode-media shortcode-media-youtube"> <span class="rm-shortcode" data-rm-shortcode-id="aab663db13dc70357bf55fd19cf3cbd3" style="display:block;position:relative;padding-top:56.25%;"><iframe frameborder="0" height="auto" lazy-loadable="true" scrolling="no" src="https://www.youtube.com/embed/P54xp4IUxno?rel=0" style="position:absolute;top:0;left:0;width:100%;height:100%;" width="100%"></iframe></span> <small class="image-media media-caption" placeholder="Add Photo Caption...">Video 1. VicOne researchers demonstrate a vulnerability that prevents a robot’s remote-stop command from working.</small> <small class="image-media media-photo-credit" placeholder="Add Photo Credit...">VicOne</small> </p><h2>A published patch still needs to be proven safe<strong></strong></h2><p>AI can shorten the search for weaknesses, but a robotics team still needs to determine whether a fix is safe to deploy. Even after a software maintainer releases a patch, teams must identify affected systems, check compatibility, and verify that the update preserves intended behavior.</p><p>An earlier successful test remains useful evidence. However, a newly discovered weakness may reveal conditions that test never examined.</p><p>Three questions can help teams respond:</p><ul><li><strong>Where do we use the affected software?</strong> Include training computers, simulation environments, and deployed robots.</li><li><strong>What could the weakness change?</strong> Assess whether it could affect models, test results, operator visibility, or robot behavior.</li><li><strong>How do we verify protection?</strong> Confirm that the fix addresses the weakness and that the updated system still operates as intended.</li></ul><p>When immediate updating is impractical, teams may need temporary safeguards, such as restricting access to an affected service, while a validated fix is prepared.</p><h2>From discovery to safer deployment</h2><p>As details from the 28 findings become available through coordinated disclosure, the next step is for robotics manufacturers is to connect each finding to its potential effect on development, deployment, and robot behavior.</p><p>That work also requires understanding how threats could affect robot behavior and how to protect systems already in operation. <a href="https://vicone.com/products/radeis/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" target="_blank"><span>VicOne Radeis</span></a> helps teams identify potential zero-day vulnerabilities, assess AI model and system security, and use simulation to evaluate how these risks could affect robot behavior before deployment. <a href="https://vicone.com/products/rthena/?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" target="_blank"><span>VicOne Rthena</span></a> supports threat detection and responses within approved policies during operation.</p><p class="shortcode-media shortcode-media-rebelmouse-image"> <img alt="Diagram of AI robot security: development, simulation, and operational protection steps" class="rm-shortcode" data-rm-shortcode-id="8f708a0e29ece1a698fa1ad89317b8c2" data-rm-shortcode-name="rebelmouse-image" id="36617" loading="lazy" src="https://spectrum.ieee.org/media-library/diagram-of-ai-robot-security-development-simulation-and-operational-protection-steps.jpg?id=68562257&width=980"/> <small class="image-media media-caption" placeholder="Add Photo Caption...">Figure 2. Cybersecurity strengthens robot safety evidence throughout the lifecycle. VicOne helps teams test safety limits against manipulation before deployment and monitor the conditions behind safe behavior in operation.</small><small class="image-media media-photo-credit" placeholder="Add Photo Credit...">VicOne</small></p><p><strong>Faster vulnerability discovery matters when teams can turn findings into action:</strong> identify what needs attention, verify that safeguards still hold, and confirm that a fix does not introduce new risks. That is how cybersecurity strengthens robot safety as systems continue to change.</p><h3></h3><br/><p>For a deeper look how cybersecurity can affect robot safety, download our white paper, “<a href="https://info.vicone.com/why-robot-safety-depends-on-cybersecurity-vicone?utm_source=ieee-spectrum&utm_medium=sponsored-content&utm_campaign=2026-11" rel="noopener noreferrer" target="_blank">Why Robot Safety Depends on Cybersecurity: Protecting the Information That Safety Decisions Rely On</a>.”</p> Reference: https://ift.tt/owSnW4J

Friday, November 13, 2026

How Microcredentials Help Engineers Build Trustworthy AI


<img src="https://spectrum.ieee.org/media-library/conceptual-illustration-of-planet-earth-surrounded-by-educational-objects-such-as-a-microscope-beaker-and-textbook.jpg?id=68675346&width=1200&height=400&coordinates=0%2C1042%2C0%2C1042"/><br/><br/><p>Engineering teams are moving at breakneck speed to integrate artificial intelligence, but they are running into a formidable new barrier: compliance issues. Translating high-level <a href="https://spectrum.ieee.org/ieee-global-initiative-ethical-design-ai-and-autonomous-systems" target="_self">AI ethics</a>, <a href="https://spectrum.ieee.org/6-guidelines-governing-ai" target="_self">governance</a>, and auditing into daily operational workflows has become one of the industry’s toughest challenges.</p><p>“AI ethics” refers to the principles developers use to govern and operate workflows. “Governance” refers to the rules and procedures an organization establishes for AI workflows and how they interact with organizational data. Auditing is the process of assessing and validating systems to ensure safety, fairness, transparency, and legal compliance.</p><p>Technology leads and systems engineers are <a href="https://spectrum.ieee.org/systems-thinking-vital-in-ai" target="_self">responsible</a> for adhering to a structured set of guidelines for standards and processes usually issued by a governing body such as the <a href="https://artificialintelligenceact.eu/" rel="noopener noreferrer" target="_blank">European Union</a>, the U.S. <a href="https://www.nist.gov/itl/ai-risk-management-framework" rel="noopener noreferrer" target="_blank">National Institute of Standards and Technology</a>, and the <a href="https://www.iso.org/sectors/it-technologies/ai" rel="noopener noreferrer" target="_blank">International Organization for Standardization</a>.</p><p>The regulations aim to mitigate issues such as <a href="https://spectrum.ieee.org/facial-recognition-gone-wrong" target="_self">systemic bias</a>, security vulnerabilities, and transparency risks, but they also highlight a critical skills gap.</p><p>Most traditional computer science, software, and electrical engineering degree programs were built around technical design and system architecture. They do not yet include training in AI auditing, data tracking, or safety protocols for how AI systems interact with people and other technologies.</p><p>Expecting mid-career engineers to leave their job to attend a full-time two- or four-year program to learn about the topics is not feasible. Yet treating safety and governance as an afterthought can lead to costly product delays, legal exposure, and broken public trust.</p><p>Targeted microcredential programs can bridge the knowledge gap by helping engineers build skills in compliance, AI ethics, and governance.</p><h2>The benefits of microcredentials</h2><p><a href="https://spectrum.ieee.org/microcredentials" target="_self">Skills-based microcredentials</a> represent the mastery of expertise gained through a learning program. To earn them, learners must demonstrate the skill through a validation assessment. The microcredentials can be stacked, showing a growing skill set in a defined area and allowing learners to earn several microcredentials in one program.</p><p>Microcredentials also allow engineers to upskill and master critical competencies, such as bias mitigation, algorithmic transparency, and risk auditing, without interrupting their career. Through rigorous skills-validation assessments, microcredentials give tech leaders verifiable proof that their teams can audit systems effectively before deployment.</p><p>For engineering organizations, microcredentialing creates an agile upskilling pathway. It connects software engineering and hardware design directly to regulatory frameworks, empowering teams to certify systems for safety and compliance within their development life cycles.</p><h2>Translating standards into microcredentials</h2><p>Bridging <a href="https://spectrum.ieee.org/two-new-ai-ethics-certifications" target="_self">technical capability and ethical responsibility</a> requires a trusted foundation. IEEE has long been a pioneer in <a href="https://standards.ieee.org" rel="noopener noreferrer" target="_blank">socio-technical standards</a>, most notably through the <a href="https://standards.ieee.org/initiatives/autonomous-intelligence-systems/standards/#p7000" rel="noopener noreferrer" target="_blank">IEEE 7000 series</a> for ethics in AI system design, a framework created to address ethical considerations during the development life cycle.</p><p>The <a href="https://credentials.ieee.org/" rel="noopener noreferrer" target="_blank">IEEE Credentialing Program</a> translates the organization’s world-class standards into actionable, industry-aligned microcredentials. For more than three decades, IEEE has set the gold standard for technical validation and professional certification. By pairing cutting-edge AI governance standards with targeted microcredential programs, IEEE provides engineers and IT leaders with clear, standardized road maps to build systems that are not just technically advanced but also safe, compliant, and worthy of public trust.</p><p>To learn more about the credentialing program, explore the <a href="https://credentials.ieee.org/" rel="noopener noreferrer" target="_blank">offerings</a> or <a href="https://credentials.ieee.org/contact/" rel="noopener noreferrer" target="_blank">connect with an IEEE expert</a>.</p> Reference: https://ift.tt/c9XSxPU

Windows 0-day dropped by disgruntled researcher hands hackers System rights


<p>Nightmare Eclipse, the pseudonym of a researcher with an ongoing, very public beef with Microsoft, has dropped yet another Windows zero-day.</p> <p>On Tuesday, Nightmare Eclipse disclosed SignDrop, a vulnerability that gives local users full System rights on some versions of Windows. The vulnerability resides in the <a href="https://learn.microsoft.com/en-us/windows-hardware/drivers/kernel/introduction-to-the-common-log-file-system">Common Log File System</a>, which Microsoft describes as a “general-purpose logging service used by software clients running in user-mode or kernel-mode.” By forcing a privileged process such as Windows Defender to open a crafted CLFS container, an attacker with ordinary user rights can abuse the privileged file creation to elevate privileges to System.</p> <h2>Turning a Windows defense against itself</h2> <p>The attack exploits the interaction of CLFS, Windows Defender, and Windows file permissions. The exploit uses Windows Defender to open a log file created by a standard user by calling ServerMpUpdateEngineSignature, a legitimate Defender function for processing antivirus signature updates.</p><p><a href="https://arstechnica.com/security/2026/11/windows-0-day-dropped-by-disgruntled-researcher-hands-hackers-system-rights/">Read full article</a></p> <p><a href="https://arstechnica.com/security/2026/11/windows-0-day-dropped-by-disgruntled-researcher-hands-hackers-system-rights/#comments">Comments</a></p> Reference : https://ift.tt/5J2fwVc

Community-Focused Academic Brings AI Tools to Colorado Farmers

<img src="https://spectrum.ieee.org/media-library/a-young-asian-woman-in-a-suit-jacket-smiling-while-holding-up-a-digital-tablet-wi...