Results for “Agent”

9to5Mac Michael Burkhardt SI

macOS 27.3 could change how you grant AI agents Full Disk Access, here’s how

Last month, Apple announced it’d be tightening privacy controls in a future macOS update, in response to AI agents requesting Full Disk Access then going rogue with users data without them being fully aware. Now, we have an idea of how the…

Read at 9to5Mac

arXiv (research papers) Sunghyeun Woo

Same Error, Different Recovery: How Delivery Changes LLM Agent Behavior

A tool-using language model must decide whether to retry a failed call, correct it, or stop and tell the user. The error text alone may not be enough: the same feedback can lead to different actions depending on where it appears in the conversation. We isolate this delivery problem by varying placement, guidance, retained history and turn structure. Two studies with plans fixed before data collection, across 13 open models from 3B to 70B parameters, compare the next action…

Read at arXiv (research papers)

arXiv (research papers) Mustafa S. Aljumaily SI

Treating the Model as an Untrusted Component: Deterministic Policy Enforcement, Human Approval and Taint Tracking Against Prompt Injection in Tool-Using Agents, with Validation on Open-Weights Models

Tool-using LLM agents act on instructions embedded in the data they read, so prompt injection is a structural rather than a prompt-wording problem. We compare in-prompt defences (hardened prompts, spotlighting, LLM detectors) with deterministic enforcement outside the model (four reference monitors of increasing strength), human approval and taint tracking, judged by one effect-based oracle independent of every defence in an executable environment of 13 tasks and 12 tools…

Read at arXiv (research papers)

The Verge Terrence O’Brien SI

Apple is racing to make Siri more agentic

Bloomberg's Mark Gurman says CEO John Ternus is pressuring AI czar Craig Federighi to accelerate the Siri roadmap and deliver a powerful AI agent that can compete with OpenAI's Dots or Meta's Muse. While Federighi has delivered a…

Read at The Verge

arXiv (research papers) Semil Eminovic Tech policy

Safeguarding Agentic AI in Radiology Workflows: A Multimodel Evaluation of Rule-based and Large Language Model-based Verification

Purpose: To estimate how often tool-using large language model (LLM) agents record clinically relevant false or unsupported information in simulated radiology documentation and how much safeguards applied before each action reduce these errors. Materials and Methods: In this simulation study (October 2026), 48 synthetic, radiologist-reviewed base scenarios covered finding communication, tumor board preparation and follow-up management. Forty-three also had a…

Read at arXiv (research papers)

TechCrunch Connie Loizos Security

He arrested a hacker, then spent 18 months clearing his own name. Now he’s a cybersecurity VC.

Stephen Ward, a former Secret Service agent and Home Depot CISO, is preparing a second fund of roughly $250 million for Brightmind Partners.

Read at TechCrunch

arXiv (research papers) Brijesh Dongol SI

On the Relative Security of Prompt Injection: Mechanised Definitions, Defences and Verdicts for LLM Agents

An LLM agent that reads untrusted content, such as an email or a web page, may follow instructions hidden in it. Noninterference is too strong for such agents, since a useful agent lets untrusted content influence what it does and may disclose the user's data by design, and declassification needs a release policy written for each task. We instead state security relative to a reference agent, which runs the same plan under the same runtime monitor with a model that never…

Read at arXiv (research papers)

The New Stack Rashmi Erande

Scaling an autonomous agent from one team to fifty: A serverless-to-container decision framework

QA teams shipping complex products generate test failures around the clock. Every failure needs a bug ticket with steps, screenshots,

Read at The New Stack

9to5Mac Michael Burkhardt Gadgets

Apple aiming to make Siri ‘more like an agent’ as soon as next year: report

Apple finally shipped its long awaited personalized Siri with iOS 27. The company lived up to its promises, and delivered on its hefty goals that it set out back in 2024. That said, Siri still lacks in terms of proactively taking actions…

Read at 9to5Mac

The New Stack Amanda Caswell

Perplexity’s AI agents keep working after you close the tab

Perplexity’s new Teammates can coordinate multiple agents, remember where previous sessions left off, and keep working without another prompt. The

Read at The New Stack

arXiv (research papers) Yiming Sun SI

When Coding Agents Get Arms: Safety Risks Across the Digital-Physical Boundary

Advances in spatial understanding have enabled frontier models to power coding agents that control a real robot arm zero-shot and end-to-end, without task-specific demonstrations or retraining. This brings two new safety challenges. First, the extensive knowledge of frontier models lets these agents perform tasks that earlier VLA and WAM paradigms could not, opening the door to new forms of capability misuse. Second, coding agents can access digital resources such as web…

Read at arXiv (research papers)

SecurityWeek Ionut Arghire Security

Brava Security Emerges From Stealth With $15 Million in Funding

Its multi-agent AI platform generates the telemetry an attack would produce to fix gaps in enterprise defenses.

Read at SecurityWeek

arXiv (research papers) Doron Shavit SI

Least-Privilege Auditing of Agent Tool Calls with Intent Graphs and Reasoning-Guided Escalation

LLM agents that call tools are exposed to indirect prompt injection: instructions hidden in tool outputs can steer an agent into calls the user never asked for. We present a runtime audit that derives an executable intent graph from the user's objective and the deployment's tool catalog with one LLM call per task, and checks every tool call against it with a deterministic loop that makes no LLM call (median 5.7 microseconds per call). Two further mechanisms recover utility…

Read at arXiv (research papers)

MarkTechPost Asif Razzaq SI

Google AI Releases Elephant-Goldfish Model Plugin for Spec-First Coding

Google's Elephant-Goldfish Model plugin makes coding agents draft specs, then runs 5 zero-context review stages before any code.

Read at MarkTechPost

The Decoder Jonathan Kemper SI

AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test

Alibaba's new coding-agent benchmark, Co-Design Bench, tests AI agents on real user requests that already tripped them up in practice. Even GPT-6 Astra, the top performer, fully solves only 26.4 percent of tasks. On older benchmarks, the same models score above 99 percent. The article AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test appeared first on The Decoder.

Read at The Decoder

arXiv (research papers) Riya Bharathwaj SI

ErosionBench: Probing Alignment Decay in Enterprise Multi-Agent Systems

As enterprises delegate consequential work to multi-agent LLM systems, benign-looking requests can escalate into unsafe actions as they propagate across sessions through poisoned shared state, long-context documents, and trusted peer messages: no single message is malicious, yet the assembled tool action is. Prior benchmarks isolate a single agent under prompt injection or score multi-agent attacks by their final output, leaving cross-session erosion of shared state out of…

Read at arXiv (research papers)

arXiv (research papers) Zhangmai Li

ICLayout-Bench: Benchmarking LLM Agents for Custom Circuit Layout Design

Large language model (LLM) agents are being explored for circuit design, but their ability to produce high-quality custom layouts has not been extensively evaluated. ICLayout-Bench is a benchmark in which agents use electronic design automation (EDA) tools to construct and optimize GDSII layouts from transistor-level circuit netlists. These layouts are evaluated using physical verification, parasitic extraction, and post-layout simulation. Layouts that pass the physical and…

Read at arXiv (research papers)

arXiv (research papers) Bibek Bhandari SI

Do Benign-Trained Hidden-State Probes Detect Tool Errors Under Adversarial Conditions?

Although hidden-state probes efficiently monitor model operation, it remains unclear whether probes trained on benign tool use remain useful for detecting incorrect actions under adversarial conditions. In this research, we assess this problem from the perspective of operational tool-action correct- ness. we use four AgentDojo suites, four conditions, and three models: Qwen2.5-7B-Instruct, Qwen2.5-14B-Instruct, and GLM- 4.6V-Flash (9B). We perform a five-fold, task-grouped…

Read at arXiv (research papers)

Journal articles (OpenAlex) Susan Beth Harper, Μ. D. Salman · Agriculture

Biosecurity in Laboratory Animal Facilities Supporting Agricultural Research: Concepts and Practical Applications

Biosecurity in laboratory animal facilities supporting agricultural research is often viewed as a means of preventing the introduction, transmission, or release of infectious agents. These functions remain essential in facilities housing livestock, poultry, aquaculture species, and laboratory animal models used to investigate agricultural priorities. Such research supports the development and evaluation of veterinary medicines, diagnostic methods, feed and nutrition…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Siqi Sun, Hualong Yu, Boyin Jin et al. · Machines

PC-MAPPO: Phase-Conditioned Rollout Buffering for Multi-Robot Collaborative Motion Planning

Phase-dependent experience creates a practical tension between formation maintenance and obstacle-aware navigation in multi-robot motion planning. This paper presents PC-MAPPO, a phase-conditioned rollout framework for multi-robot formation and obstacle-aware navigation in the evaluated simulation setting. The method retains the Multi-Agent Proximal Policy Optimization (MAPPO) optimizer and reorganizes on-policy data through separate formation and navigation buffers…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) Jie Ke · Journal of Leadership Accountability and Ethics Tech policy

Before the Policy: Developing Academic Leaders and Institutional Capacity for the Governance of Agentic AI

Institutional policies for agentic artificial intelligence are only as good as the leaders and structures that must operate them, yet presidents, provosts, deans, and boards receive almost no systematic preparation for AI governance. This conceptual paper defines the institutional layer of agentic AI governance, explains why senior leaders are its least developed actors, and proposes an institutional AI governance capacity model with three layers: individual governance…

Read at Journal articles (OpenAlex)

Journal articles (OpenAlex) V. Ramalingam, V. S. Lee, Yew Hoong Wong · Materialwissenschaft und Werkstofftechnik

Silane coupling agent for semiconductor manufacturing

Silane coupling agents (SCAs) have evolved from conventional adhesion promoters into multifunctional interfacial materials essential for semiconductor packaging reliability. As device miniaturization, heterogeneous integration, and advanced packaging technologies intensify interfacial challenges, SCAs increasingly contribute to moisture resistance, stress mitigation, electrical performance, and thermal stability. This review examines six decades of SCA development, from…

Read at Journal articles (OpenAlex)

Axios Ben Berkowitz

"Congress will investigate": Nvidia CEO Jensen Huang faces probe over alleged Chinese spy ties

Nvidia CEO Jensen Huang is facing increasing congressional scrutiny after multiple reports Friday night suggested he had a romantic relationship with a purported Chinese spy. Why it matters: Huang is arguably the world's most powerful CEO, the man who runs one of the world's largest companies and whose products are central to the world's most important technology. The mere suggestion that he was in some way compromised by a Chinese intelligence agent could quickly become a…

Read at Axios

GeekWire Todd Bishop SI

JPMorgan and Mastercard back Seattle startup Firmly’s bet on AI shopping agents

Firmly raised $15 million to expand software that lets online retailers sell through AI shopping assistants, marketplaces and other channels without building a separate integration for each one. The company also launched Ask Agent, an AI…

Read at GeekWire

The Decoder Manuel Uth SI

AI auditors catch most agent misbehavior but flood reviewers with false alarms, new study finds

AI models can often spot problematic behavior by other AI agents in logs, according to a study by Anthropic researcher Andy Hall. GPT-5 found about 98 percent of reward-hacking cases, but the more exhaustively the search is tuned, the more false alarms pile up. That's not yet enough for reliable safety checks. The article AI auditors catch most agent misbehavior but flood reviewers with false alarms, new study finds appeared first on The Decoder.

Read at The Decoder

arXiv (research papers) Thanh M. Truong-Le SI

Agents of Chaos: A Systematization of Knowledge (SoK) on Malicious Attacks on Agentic AI Systems

Agentic AI systems extend large language models with tools, memory, retrieval, browser or computer-use interfaces, code execution, external services, and multi-agent coordination. Although these capabilities enable agents to act beyond text generation and perform tasks, they also create attack paths through which adversaries can exploit untrusted content, deceptive metadata, poisoned state, compromised skills, mutable dependencies, or malicious code. Such attacks can…

Read at arXiv (research papers)

The New Stack Adrian Bridgwater

Why AWS wants developers to stop babysitting their AI agents

AWS first introduced Kiro, its agentic, spec-driven development environment, in July 2025. It also introduced Kiro Crew in August of

Read at The New Stack

The Decoder Jonathan Kemper SI

Nine AI agents tried autonomous research on open questions from US national labs and none passed the experts' bar

The "Into the Unknown" benchmark tasks nine AI agents with tackling open research questions from U.S. national laboratories. None of the 135 submissions meets all requirements. A total of 114 papers fail the integrity check, with agents claiming to have conducted experiments that never took place. GPT-6 Astra performs best but still doesn't meet the experts' standards. The article Nine AI agents tried autonomous research on open questions from US national labs and none…

Read at The Decoder

arXiv (research papers) Yihuang Kang Tech policy

Towards Identity-Preserving LLM Agents under Model Transition

When an LLM agent's backend model is upgraded or replaced, operators typically reload the same persona, memory, and rules onto the new model and assume the agent will behave as before. We show this assumption requires empirical validation-and that whether identity is preserved depends on the specific model pairing. Across 5,760 compliance-testing runs in 12 regulated domains with three open-weight models, we find that some transitions produce significant behavioral drift…

Read at arXiv (research papers)

Tech in Asia Grace Priscilla Teo Startups

Meta adds Spaces to Muse AI agents for organizing projects

Zuckerberg said Muse had 10 million US and Canada users less than two months after launch.

Read at Tech in Asia

Older stories