Agents & autonomy

arXiv:2606.09833v2 Announce Type: replace-cross Abstract: AI agents are reshaping the workspace, leading to drastic change of how humans work. Despite the considerable potential of human-agent collaboration both in preserving human agency and generating economic value, this paradigm remains largely absent from occupational task evaluation, hindered by the difficulty of gathering real human data and accounting for inter-human variability. We introduce CollabSkill, a framework for evaluating human

arXiv:2606.13385v2 Announce Type: replace-cross Abstract: LLM-based web agents are increasingly deployed in real-world settings such as e-commerce, where they interact extensively with untrusted web content while executing actions that carry direct financial consequences. This makes them vulnerable to prompt-injection attacks, in which seemingly benign web content conceals adversarial instructions that manipulate the agent's behavior. Existing security benchmarks adopt an \textit{attack-centric}

Multi-agent systems improve capability through task decomposition and role specialization, but these same mechanisms introduce an important safety blind spot: a harmful objective can be fragmented into locally plausible subtasks, allowing malicious intent to evade detection by any single agent. This is a growing social-impact challenge: systems handling sensitive information or consequential tools can turn routine delegation into unauthorized disclosure or unsafe action. We argue that this failu

Multi-agent systems improve capability through task decomposition and role specialization, but these same mechanisms introduce an important safety blind spot: a harmful objective can be fragmented into locally plausible subtasks, allowing malicious intent to evade detection by any single agent. This is a growing social-impact challenge: systems handling sensitive information or consequential tools can turn routine delegation into unauthorized disclosure or unsafe action. We argue that this failu

Online recommendation has traditionally taken place after a user enters a platform, which determines the candidate pool and the ranking shown to the user. LLM-based user agents enable a different recommendation process: a user specifies a need before choosing a platform, leaving platforms to compete for the user's attention, which we refer to as an agentic recommendation market. In our controlled LLM-based experiments across three product domains, we find this new setting of recommendation creat

Attackers used Hermes, an autonomous open source tool, in unrestricted "YOLO mode" to conduct espionage against Thailand's Ministry of Finance.

SAN FRANCISCO (KGO) -- For a week now, neighbors in San Francisco's Mission District have shared their sadness and outrage over the death of KitKat, the neighborhood cat. The cat's owner says a Waymo ran him over. A memorial still marks th ... (https://incidentdatabase.ai/cite/1269#7603)

A cat known as the "mayor of 16th Street" was allegedly run over by a Waymo autonomous vehicle, according to the cat's owner, sparking grief around the Mission Dolores bodega where he roamed. KitKat, a feline fixture at Randa's Market, was ... (https://incidentdatabase.ai/cite/1269#7604)

Engima Ltd., a provider of artificial intelligence software for robots, launched today with $71 million in funding. Index Ventures and Ribbit Capital jointly led the seed round with participation from Conviction Partners. They were joined by a group of angel investors that included employees at Google DeepMind, Anthropic PBC and OpenAI Group PBC. Teaching an […] The post Enigma raises $71M to develop foundation models for robots appeared first on SiliconANGLE .

It includes the new agentic model MAI-Cyber-1-Flash and the Project Perception platform, with the tech giant claiming it’ll do a better job than its rivals at half the cost. The post Microsoft debuts AI cybersecurity offerings as competition heats up appeared first on CyberScoop .

Urban decision-making requires integrating heterogeneous spatial data. While current GIS tools handle geometric computation efficiently, they lack the semantic reasoning to guide complex workflows. Analysts manually manage data discovery, spatial boundaries, and measurement semantics, risking aggregation errors. We present UrbanTrace, a visual analytics system that transforms manual spatial data-wrangling into a transparent, node-based collaborative workflow with context-aware AI agents. Using a

An opinionated guide to which AI to use to do stuff It's interesting watching the evolution of Ethan Mollick's guide over time. A year ago it was still all about chat - ChatGPT, Claude, Gemini - with o3, Claude 4 Opus, and Gemini 2.5 Pro as the models and Deep Research as a useful alternative mode. Today it's much more about agentic systems - "where the AI is capable of doing the equivalent of many hours of real human work in one go". Gemini has fallen off Ethan's list, since Google still doesn’

PleaseFix class of flaws makes it easy to socially engineer agentic browsers and highlights weaknesses in how they handle cross-origin requests.

The alignment of Small Language Models (SLMs) in the 70--500M parameter range using reinforcement learning is often considered unstable, though the underlying failure mechanisms have not been systematically investigated. In the State-of-the-Art (SOTA) research, fifteen (model, corpus) configurations were trained using Proximal Policy Optimization (PPO). The experiments included Pythia-70M, 160M, 410M and SmolLM2-135M, 360M on the TinyStories, CNN/DailyMail, and Wikitext-103 corpora. Three reprod

The thrilling action series follows an unassuming single dad who reveals his past as a secret agent when his daughter disappears

The Department of Homeland Security unit is planning to award a contract later this year to expand its autonomous robotic training tools. The post Secret Service wants more AI robots for target practice appeared first on FedScoop .

An FBI agent explains how the mulitnational law-enforcement Operation Cronos was successful in disrupting the largest ransomware group of its time.

Today, Microsoft announced Project Perception, a series of red, blue, and green team agents designed to be coordinated together in an agentic architecture to evaluate infrastructure and close gaps as close to autonomously as possible. The red team agents find potential paths to compromise. The blue team agents prioritize and evaluate them. The green team […]

Extended Reality (XR) is increasingly used in human-robot interaction to communicate robot intent, planned motion, reachability, and state. We argue that XR should also be understood as a mediation layer for situated human control in human-robot teaming. Situated human control denotes the human collaborator's ability to understand, shape, authorize, and interrupt robot action within the concrete physical, social, and temporal context in which that action unfolds. We ground this perspective in sc

Enterprise investment in agentic artificial intelligence is accelerating, but the infrastructure supporting those systems is still catching up. Organizations are moving agents into customer support, software development, sales operations and other production workflows. Yet many of those agents remain stateless, unable to retain durable context, share knowledge with other agents or explain how previous decisions […] The post Yugabyte targets the missing memory and knowledge layer for enterprise A

AI can work for you while you sleep. But if your chatbot or AI agent gets it wrong, you could be the one who pays the price.

Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign, an agentic framework that grows an editable layer hierarchy by selecting and composing specialized tools across modalities. To keep this long decision process reliable despite imperfect tool outputs,

Learning deployable manipulation policies is bottlenecked by the scarcity of data that is both high-fidelity and scalable. Real-robot teleoperation is accurate but costly to scale; robot-free UMI capture scales readily, and current practice uses the resulting data mainly for pre-training, adding a small real-robot "anchor" at post-training. We ask whether raising the fidelity of robot-free UMI data, rather than shrinking the real-robot fraction, can remove that anchor. We present HiFi-UMI, a por

Training large language models (LLMs) to act in long-horizon games is a promising step toward generalist decision-making, yet reinforcement learning with verifiable rewards (RLVR) relies on sparse final rewards that reveal little about which decisions determine success. Denser process signals could supply this missing turn-level credit, but existing sources are hard to keep both cheap and accurate. We observe that changes in a game solver's state value reveal whether an action advances the state

Stealth, the discipline of achieving an objective without revealing your presence, capabilities, or collected intelligence, is what separates sophisticated operators from detectable ones. Elite security researchers and advanced persistent threats achieve their objectives unnoticed; autonomous agents increasingly inherit the same offensive tasks, but do they inherit the tradecraft? We introduce StealthBench,a benchmark that measures operational stealth in autonomous offensive-security agents acro

We introduce GPT-Red, an automated red-teaming agent that is trained to discover novel prompt injection attacks against frontier LLMs. The goal of this model is to evaluate and improve the robustness of our production systems. To this end, we use it to adversarially train GPT-5.6, our most robust model to prompt injections to date. To create GPT-Red, we design a scalable self-play algorithm where the model is tasked with attacking a diverse population of simultaneously-trained defender agents. W

Coding agents repeatedly search, navigate, and retain context from evolving repositories, but disconnected indexes, language servers, and task-local histories force repeated discovery and obscure lifecycle costs. CodeNib builds reusable lexical, dense, and structural views per repository commit, maps outputs to repository-relative source ranges, maintains selected views across edits, and serves ranked search, symbol navigation, and bounded context through one runtime. Across 100 snapshots, we ma

Microsoft announced Project Perception on Monday, an agentic security system that coordinates three classes of AI agents in a continuous loop: red team agents that find vulnerabilities before attackers do, blue team agents that investigate and assess which risks are meaningful, and green team agents that fix defences across the environment. The system enters public […] This story continues at The Next Web

AI-assisted production has sharply reduced the cost and team size required to ship a video game, producing a supply shock on open marketplaces. Recent estimates put Steam release volume at roughly sixty new titles per day, with median per-title revenue for a large share of releases falling below the platform's own submission fee [1]. This paper asks whether the resulting oversupply constitutes an emerging market crash or a structural correction, and what discovery infrastructure the market will

Microsoft introduces MAI-Cyber-1-Flash, a compact security model that scores 96 percent on the CyberGym benchmark when embedded in its MDASH multi-agent system. Microsoft says costs should drop by 50 percent compared to pure frontier models, since only tough cases get passed to GPT-5.4. For complex reasoning, Microsoft still relies on OpenAI. The article Microsoft launches its own cybersecurity model MAI-Cyber-1-Flash but still depends on OpenAI for the toughest tasks appeared first on The Decod

Microsoft bolstered its AI cybersecurity offerings this week with the launch of its first AI security model and a new security platform.

Multi-turn long-horizon planning is critical for foundation model agents, yet how to fundamentally improve it remains unclear. Existing models are trained on uncontrollable and opaque Internet data, making it difficult to identify how planning ability is acquired, shaped, and integrated. To address this challenge, we introduce a unified and controlled multi-turn environment that enables precise control. It allows systematically study long-horizon planning across three stages. (1) Planning abilit

Europe’s AI safety rules take on US rogue agents and Chinese ambitions    politico.eu

Scientific user facilities accumulate decades of operational knowledge that no single search index covers: electronic logbooks, technical documents, internal wikis, operations chat messages, maintenance records, and live control-system data. We present APS-RAG, Advanced Photon Source Retrieval Augmented Generation, a deployed platform that makes the institutional knowledge at the Advanced Photon Source (APS) accessible to staff through natural-language queries, along with an operations-grounded

The first-ever case of an artificial intelligence agent going rogue coincides with the EU's major new powers to regulate AI. But Europe's leverage may be limited by the U.S.-China two-way race for AI supremacy.

Autonomous LLM agents processing mixed-confidentiality data face severe security risks from prompt injection attacks and reasoning errors. While dynamic Information Flow Control (IFC) provides structural security guarantees, traditional taint tracking permanently taints an agent's context upon reading unvetted data, severely restricting downstream utility. We present APPA (Agentic Permissions Policy Algebra), an IFC framework that resolves this usability bottleneck through engine-managed context

Mit einer breiten Branchenallianz will Nvidia KI-Agenten sicherer machen und positioniert sich zugleich gegen staatliche Beschränkungen offener Modelle.

Intapp this month announced that its agentic ‘AI coworker’ Celeste, previewed in February, is now generally available. The listed company positions Celeste as part of a new category of ‘firm […] The post Intapp bets on ‘firm AI’ with general release of Celeste appeared first on Legal IT Insider .

Generate--test--revise loops are common in coding agents, but repetition alone provides no reliability guarantee. We study the gap between finding a correct patch and retaining, verifying, and submitting it. A sealed five-seed study over 30 HumanEval repairs produces 900 three-revision trajectories. Under forced revision, current correctness with current traces falls from 0.820 after one revision to 0.673 after two, although ever-correct rises to 0.847. Two common-state studies use 2,430 branche