{"data":[{"id":"4642ae78-30a1-49f0-b6a8-d378aaa80145","title":"TMT: Runtime Backdoor Detection for Vision-Language-Action Policies on Unseen Tasks","summary":"TMT is a runtime backdoor detector for vision-language-action (VLA) policies, built on token manifold and latent transition modeling and trained on benign rollouts. The authors also explore policy purification through self-distillation, using a frozen copy of the backdoored policy as a teacher. In a post-hoc comparison with ten baselines, TMT achieves state-of-the-art detection on unseen tasks across three VLA backdoor attacks.","sourceUrl":"https://arxiv.org/abs/2610.09462v1","publishedAt":"2026-10-07T05:20:16.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["security","research"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":[],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":["other"],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-10-07T05:20:16.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["integrity","safety"],"aiComponentTargeted":"model","llmSpecific":false,"classifierConfidence":0.95,"researchCategory":"preprint","atlasIds":null},{"id":"00db42bd-443e-42b4-8fba-5be6e89498b4","title":"Lie Rarely, Lie Big: Stealthy Insider Attacks on LLM Robot Teams","summary":"Researchers study how a single compromised robot can corrupt the shared map of an LLM-agent multi-robot survey team, using a grounded task where measurements can be checked against the physical world at a cost to mission budget. They derive a lower bound on the probability that the adversary's reports are verified and an upper bound on the map error any stealthy adversary can cause, with the worst attack switching from rare full-magnitude lies to small hidden biases above a critical verification level. Experiments with LLM-agent honest robots show both bounds hold at the budget the attack spent, and that LLM robots re-check records without bias in which records they check but unpredictably in how much.","sourceUrl":"https://arxiv.org/abs/2610.04744v1","publishedAt":"2026-10-03T20:20:57.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["security","research"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["LLM agents"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":["other"],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-10-03T20:20:57.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["integrity"],"aiComponentTargeted":"agent","llmSpecific":true,"classifierConfidence":0.9,"researchCategory":"preprint","atlasIds":null},{"id":"8fbea738-1949-4480-ba8c-ebfbbf377d18","title":"(Un)Cooperative Robots: From Compliance AI to Principled Uncooperative AI","summary":"Stuart Russell's human-compatible AI framework proposes that AI systems should defer unconditionally to human operators. This paper argues that total deference carries its own risks and limitations, and reports a HICSS workshop panel on (un)cooperative robots that can refuse human directives for principled reasons. The panel identified eleven research themes, including authority hierarchies, benevolent uncooperativeness, and AI liability, and called for research beyond compliance-centric AI design.","sourceUrl":"https://aisel.aisnet.org/thci/vol18/iss3/5","publishedAt":"2026-09-30T17:35:37.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["research","safety"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":[],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-30T17:35:37.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":["safety"],"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.85,"researchCategory":"peer_reviewed","atlasIds":null},{"id":"5e7b018a-383f-43f0-a709-8263d6a45f08","title":"Exploiting Vulnerabilities: Universal Adversarial Attacks on Vision-Language-Action Models in Robotics","summary":"Researchers propose a Universal Adversarial Object, a sphere with an optimized surface texture, that degrades the task success of Vision-Language-Action (VLA) models when placed in a robot's field of view. Their multi-level attack framework disrupts trajectory planning, task execution and action control, and was validated in simulated and real-world robotic settings. The object reduces average task success rates by 31.2% to 39.9% for two VLA models, Pi0 and RDT, with success rates dropping to near zero in complex scenarios.","sourceUrl":"https://arxiv.org/abs/2609.39178v1","publishedAt":"2026-09-30T07:34:24.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["security","research"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["Pi0","RDT"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":["model_evasion"],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-30T07:34:24.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["integrity","availability","safety"],"aiComponentTargeted":"model","llmSpecific":false,"classifierConfidence":0.95,"researchCategory":"preprint","atlasIds":null},{"id":"c494fa1d-2b9a-43fe-bb99-223e61872a06","title":"What work can robots do?","summary":"Anthropic researchers built a robot exposure index using Claude to rate how well present-day robots can perform US job tasks. They find robots can already do 74% of physical tasks, making up 34% of working hours, mostly in limited settings. Robots are cost-competitive for only 0.3% of tasks, and jobs more exposed to robots saw larger wage and employment declines over the past 50 years.","sourceUrl":"https://www.anthropic.com/research/what-work-can-robots-do","publishedAt":"2026-09-30T05:00:00.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["research","industry"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["Anthropic"],"affectedVendorsRaw":["Claude","LLMs"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-30T05:00:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":"industry","atlasIds":null},{"id":"25d266d9-8358-441c-8cb1-84ad1ef7c031","title":"Exein Secures $270M at $1.7B Valuation for Physical AI Security","summary":"IoT cybersecurity startup Exein announced a $270 million funding round led by Headline, bringing its total raised above $600 million and its valuation to $1.7 billion. The company, founded in 2018 in Italy, plans to use the money to build a proprietary foundation model for Physical AI security and to expand globally, with a focus on the US market.","sourceUrl":"https://www.securityweek.com/exein-secures-270m-at-1-7b-valuation-for-physical-ai-security/","publishedAt":"2026-09-15T15:45:13.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry","security"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["Exein","HelmGuard","Cylake","Catch","HiddenLayer"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-15T15:45:13.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":null,"atlasIds":null},{"id":"d4d71938-1864-4294-bf70-56157df5ff56","title":"Jensen Huang puts Trump on speakerphone onstage to announce robots won’t take over the world","summary":"Nvidia CEO Jensen Huang put President Trump on speakerphone during a Monday appearance at the All-In Summit, held by the All-In Podcast. On the call, Trump called recent fears about AI development a \"hoax\" and told the crowd that \"the robots will not be taking over.\"","sourceUrl":"https://www.theverge.com/tech/995079/president-donald-trump-calls-nvidia-ceo-jensen-huang-all-in-summit","publishedAt":"2026-09-14T20:03:37.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry","policy"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["NVIDIA","Anthropic"],"affectedVendorsRaw":["NVIDIA","Anthropic","Dario Amodei essay \"We Must Pace the Frontier\""],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-14T20:03:37.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":null},{"id":"de38b110-ce41-4c74-8301-0e79731abafb","title":"A Comparative Survey of Security Risks in AI Systems: From LLMs to AI Agents and Embodied Agents","summary":"This ACM Computing Surveys paper, titled \"A Comparative Survey of Security Risks in AI Systems: From LLMs to AI Agents and Embodied Agents,\" appears in Volume 58, Issue 15, pages 1-38, November 2026. The source text provided contains only the citation details, so its research question, method and findings cannot be summarized from it.","sourceUrl":"https://dlnext.acm.org/doi/abs/10.1145/3837083?af=R","publishedAt":"2026-09-03T12:01:34.509Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["research","security"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":[],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.95,"researchCategory":"peer_reviewed","atlasIds":null},{"id":"7e068804-86d3-4120-a85f-1e8af16693fb","title":"Hugging Face's new duck robot is selling fast. A Chinese chip powers it","summary":"Hugging Face's French subsidiary Pollen Robotics launched the Microduck, a 800g programmable duck-shaped robot priced at $399. It has sold more than 10,000 units, earning over $4 million, and delivery times for new orders have slipped past the initial promise of Christmas 2026. The robot runs on Rockchip's RK3566 chip, which uses technology licensed from ARM.","sourceUrl":"https://www.cnbc.com/2026/09/01/hugging-faces-new-duck-robot-is-selling-fast-a-chinese-chip-powers-it.html","publishedAt":"2026-09-01T07:30:49.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["HuggingFace","NVIDIA"],"affectedVendorsRaw":["Hugging Face","Pollen Robotics","Microduck","Rockchip","ARM","Nvidia"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-09-01T07:30:49.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":null,"atlasIds":null},{"id":"cf25fa32-161a-4f15-9469-2514d2e8c070","title":"DeepSeek looks for fresh capital as founder’s quant empire navigates China’s choppy IPO market","summary":"DeepSeek, the Chinese AI lab financed for years by founder Liang Wenfeng's hedge fund High-Flyer Quant, is now seeking outside investors to fund its growing ambitions. High-Flyer's affiliates secured pre-IPO allocations in China's memory chipmaker CXMT, worth a combined $26 million, and robot maker Unitree Robotics, worth $5.8 million, while DeepSeek itself invested in Unitree as a strategic investor. Analysts say High-Flyer's revenue has become unstable after AI and chip stock volatility caused losses in July.","sourceUrl":"https://www.cnbc.com/2026/08/28/deepseek-founder-liang-wenfeng-high-flyer-china-tech-ipos-funding.html","publishedAt":"2026-08-28T02:17:01.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["DeepSeek","High-Flyer Quant"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-08-28T02:17:01.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null},{"id":"f9d469a4-e750-46c9-9225-8f7391ebb1ec","title":"Hugging Face’s new robot is an adorable rollerskating duck","summary":"Hugging Face's Pollen Robotics has launched the Microduck, its second AI robot, a one-eyed biped just under 10 inches tall. It is available to preorder for $399 in four colors, with shipping planned \"before Christmas 2026.\"","sourceUrl":"https://www.theverge.com/gadgets/985549/hugging-face-microduck-robot","publishedAt":"2026-08-27T13:44:00.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["HuggingFace"],"affectedVendorsRaw":["Hugging Face","Pollen Robotics","Microduck"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-08-27T13:44:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":null,"atlasIds":null},{"id":"5a4ececf-6e79-4625-b233-d5fe73746a72","title":"A Comparative Survey of Security Risks in AI Systems: From LLMs to AI Agents and Embodied Agents","summary":"This ACM Computing Surveys article, published in Volume 58, Issue 15 (pages 1-38) in November 2026, surveys security risks across AI systems, from large language models to AI agents and embodied agents. The source text provided contains only the bibliographic citation and no article content, so no findings or methods can be reported from it.","sourceUrl":"https://dl.acm.org/doi/abs/10.1145/3837083?af=R","publishedAt":"2026-08-23T12:01:48.150Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["research","security"],"issueType":"research","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":[],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.95,"researchCategory":"peer_reviewed","atlasIds":null},{"id":"a21ffa02-f508-426f-96b7-5aa37000925f","title":"One of science fiction’s greatest writers warned us about a AI. Does he also hold the remedy? | Alan Finkel","summary":"Alan Finkel proposes a modern equivalent of Isaac Asimov's laws of robotics, framed as three laws of AI. He draws on Elon Musk's July prediction that AI-powered robots could dominate the physical world and might stop taking orders from people. The article notes that current government actions, including US restrictions on frontier AI models and EU AI regulations that took effect this year, fall short of guardrails that would give AI a desire for humanity to prosper.","sourceUrl":"https://www.theguardian.com/technology/commentisfree/2026/aug/07/science-fiction-warned-us-about-an-ai-powered-dystopian-future-does-it-also-hold-the-remedy","publishedAt":"2026-08-07T04:00:08.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["policy","safety"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["OpenAI","Anthropic"],"affectedVendorsRaw":["OpenAI","Anthropic","frontier AI models"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-08-07T04:00:08.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":null},{"id":"02b087b1-632f-4bc8-be01-7928689486c8","title":"Google DeepMind’s new AI model can control a robot’s entire body","summary":"Google DeepMind says the latest version of its Gemini Robotics AI model can control entire humanoid robots. The new Gemini Robotics 2 model supports whole-body motions from feet to fingertips, where its predecessor focused on upper-body control. Google shared videos of Apptronik's Apollo 2 robot bending to pick up a watering can and taking items off a shelf.","sourceUrl":"https://www.theverge.com/tech/973276/google-deepmind-gemini-robotics-2-whole-body","publishedAt":"2026-07-30T17:18:45.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["Google"],"affectedVendorsRaw":["Google DeepMind","Gemini Robotics 2","Apptronik Apollo 2"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-07-30T17:18:45.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null},{"id":"ba9bc682-86c3-4559-af70-b2c2f7b0991e","title":"Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration","summary":"Google launched Gemini Robotics ER 2, its most capable embodied reasoning model for robotics, on July 30, 2026. The model acts as a high-level planner that talks with humans, understands the physical world, and hands motor execution to lower-level vision-language-action (VLA) models, while also calling tools such as Google Search or user-defined functions. It is publicly available through the Gemini API and Google AI Studio, and in private preview on Gemini Enterprise Agent Platform.","sourceUrl":"https://deepmind.google/blog/gemini-robotics-er-2-powering-robotics-with-video-understanding-task-orchestration-and-multi-robot-collaboration/","publishedAt":"2026-07-30T15:00:59.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry","research"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["Google"],"affectedVendorsRaw":["Gemini Robotics ER 2","Gemini Robotics ER 1.6","Gemini Live API","Gemini API","Google AI Studio","Gemini Enterprise Agent Platform","Spot"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-07-30T15:00:59.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null},{"id":"e08e5402-cbaf-489f-b847-a558923b17dd","title":"Gemini Robotics 2 brings whole body intelligence to robots","summary":"Google DeepMind introduced Gemini Robotics 2, a set of three models that give robots whole-body control, dexterous manipulation and multi-robot collaboration. The vision-language-action model controls full humanoids and bi-arm robots, while Gemini Robotics ER 2 handles reasoning and planning and is available on Google AI Studio and in private preview on Gemini Enterprise Agent Platform.","sourceUrl":"https://deepmind.google/blog/gemini-robotics-2-brings-whole-body-intelligence-to-robots/","publishedAt":"2026-07-28T13:21:37.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["Google"],"affectedVendorsRaw":["Gemini Robotics 2","Gemini Robotics ER 2","Gemini Robotics On-Device 2","Gemini","Google AI Studio","Apptronik Apollo 2"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-07-28T13:21:37.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null},{"id":"1ca98f30-0547-4b29-b939-c826bef0c539","title":"Nvidia unveils new AI model and expands Japan’s physical AI ecosystem","summary":"Nvidia unveiled Cosmos 3 Edge, a world model designed to help robots and vision AI agents perceive and navigate physical environments in real time. The launch extends Nvidia's physical AI push in Japan, where CEO Jensen Huang is forming a coalition that Fujitsu, Hitachi, and Kawasaki Heavy Industries intend to join. Nvidia is also expanding into healthcare and biotech, with Tokyo-1, an AI drug discovery consortium run by Xeureka, powered by the Nvidia BioNeMo Agent Toolkit.","sourceUrl":"https://www.cnbc.com/2026/07/16/nvidia-reveals-new-ai-model-and-expands-japans-physical-ai-ecosystem.html","publishedAt":"2026-07-16T11:21:56.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["NVIDIA"],"affectedVendorsRaw":["NVIDIA","Cosmos 3 Edge","Cosmos 3","BioNeMo Agent Toolkit","Microsoft","SoftBank","Fujitsu","Hitachi","Kawasaki Heavy Industries","Tokyo-1","Xeureka"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-07-16T11:21:56.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null},{"id":"6af84c95-6075-4966-97c2-f44a1a631832","title":"Pokémon Go data trained AI that could assist military drones in war zones","summary":"A Pokémon Go data set has been used to train an AI model that could help military drones determine their location in war zones. The source text describes the game as a 2016 augmented reality mobile app that used phone cameras, and reports more than 800m downloads worldwide by 2018.","sourceUrl":"https://www.theguardian.com/technology/2026/jun/12/pokemon-go-data-trained-ai-that-could-assist-military-drones-in-war-zones","publishedAt":"2026-06-12T07:06:00.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry","privacy"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["Pokémon Go","Niantic"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-06-12T07:06:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":null,"atlasIds":null},{"id":"11d04e5a-fdd4-4270-a924-06ebdcd88c68","title":"Are robots nearing their ChatGPT moment? – podcast","summary":"A robot named Lightning beat the human half marathon world record by nearly seven minutes at Beijing's half marathon. The Chinese government has pledged more than £100bn in robotics investment over the next 20 years. The podcast asks whether robots will enter everyday life as chatbots have, with Amy Hawkins and Nathan Lepora of Bristol University discussing robots in the workforce and human-like dexterity.","sourceUrl":"https://www.theguardian.com/science/audio/2026/may/28/are-robots-nearing-their-chatgpt-moment-podcast","publishedAt":"2026-05-28T04:00:56.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":[],"affectedVendorsRaw":["Lightning (Beijing half marathon robot)"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-05-28T04:00:56.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":false,"classifierConfidence":0.8,"researchCategory":null,"atlasIds":null},{"id":"b9b7719e-5831-4be5-ae90-f2b42efaaa09","title":"Would you let robots spend your money? Google is betting on it","summary":"Google unveiled a \"Universal Cart\" at Google I/O, an AI shopping tool that works across retailers and Google products such as Gemini, with YouTube and Gmail planned later. Users add items while browsing Search or chatting with Gemini, then check out through Google, and the cart tracks prices, sends in-stock notifications, suggests discounts and flags potential issues with selections.","sourceUrl":"https://www.theverge.com/news/932927/google-io-agentic-ai-shopping-universal-cart","publishedAt":"2026-05-19T17:45:00.000Z","severity":"info","cvssSeverity":null,"cvssScore":null,"labels":["industry"],"issueType":"news","cveId":null,"cweIds":null,"affectedPackages":null,"affectedVendors":["Google"],"affectedVendorsRaw":["Google","Gemini","Google I/O","Universal Cart","YouTube","Gmail"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","solution":"N/A -- no mitigation discussed in source.","attackType":[],"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"patchAvailable":null,"disclosureDate":"2026-05-19T17:45:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":null,"llmSpecific":true,"classifierConfidence":0.9,"researchCategory":null,"atlasIds":null}],"meta":{"total":31,"limit":20,"offset":0}}