Agents & autonomy

Vision-language model (VLM) agents increasingly use tools to act on 3D scenes rather than only describe them. Existing 3D benchmarks score textual responses or single-object operations, leaving agent action on complete multi-object 3D scenes under evaluated. We present SceneActBench, a benchmark for visually conditioned action across five 3D tasks under a unified agent-environment loop. Given PNG images or sampled video frames and, where applicable, supplied 3D assets, an agent acts on a 3D envi

LLM training is shifting from manual design and annotation to interaction-driven self-evolution. However, existing self-evolutionary methods face a fundamental dilemma between task diversity and verification reliability: environment-bound methods obtain precise feedback but confine learning to narrow domains, while open-ended self-generation broadens the task space but lacks reliable verification, allowing misleading rewards to pollute the training loop. We identify agent skills as a powerful mi

Computer-use agents are usually improved by strengthening perception: better models for reading a screenshot and choosing where to click. Yet a screenshot is only a lossy rendering of the underlying program state, e.g., the files, application backends, and DOM that hold the task data. Different states can produce the same pixels, while code can inspect and modify that state directly. StateAct is a code-first, multi-agent harness built around this distinction. Its main agent works directly with p

Autistic individuals often face barriers in workplace communication, where soft skills are embedded within ongoing tasks and surrounding environment context, not in isolated verbal exchange. Recent work has introduced LLM-driven agents into VR-based communication training and proposed prompting schemas that let agents generate dialogue grounded in the VR environment and the user's hand-based interactions. Building on this work, we explore how different levels of environmental grounding influence

AI-powered assistive technologies have long supported blind and low vision (BLV) people in everyday tasks, but they are general-purpose and often fall short of meeting complex, individualized, in-situ accessibility needs. Though agentic programming tools, like GitHub Copilot, have the potential to bridge this gap by lowering the technical barriers to building personal AT using natural language, the practical applicability of this creation paradigm has been unknown. We address this knowledge gap

Large language models are increasingly used to simulate human preferences in research and practical applications, raising concerns about validation, misrepresentation, and exclusion. Co-designing agents with the people they represent is a promising way to address these concerns, but participation may also mask the problems it appears to solve. This paper explores that tension through a primarily qualitative study in which 12 participants co-designed personal preference agents in the domain of ho

AegisAI co-founders developed AI agents that quickly analyze each message as a human would, paying attention to small anomalies that even the most elaborate checklist wouldn’t catch.

The next leap in lethality will depend less on advanced weapons alone and more on software-driven munitions that can adapt, scale and earn commanders’ trust. The post How adaptive weapons and trusted autonomy are reshaping the future of fires appeared first on DefenseScoop .

Black Forest Labs has released Flux 3, a multimodal foundation model that learns from images, video, and audio and can generate video with native sound for the first time. BFL's own tests put it just ahead of market leader Seedance 2.0, though independent results aren't yet available. The company ultimately wants to build a world model and is already testing Flux 3 on robotics tasks. The article Flux 3 generates videos with native audio up to 20 seconds long, a first for Black Forest Labs appear

Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments.

Creating dynamic and physically realistic 4D worlds from natural language descriptions is both fascinating and challenging. Traditional computer graphics methods rely on manual creation, requiring extensive human effort to fine-tune materials, motions, and visual fidelity. Recent advances in generative foundation models have sparked interest in learning to generate such 4D worlds from large-scale data; however, existing methods still struggle to ensure physical plausibility and controllability.

Even a current high-capability LLM can appear safer when shown a dangerous objective directly than when other agents transform and relay its direction. Using OpenAI's gpt-5.6-sol model alias, we test 25 pre-specified mirrored trade-off profiles. Direct exposure to an objective authorizing concealment, fabrication, and pressure produced advice net opposed to its target. After an Id and Censor transformed the same objective into affect and a constraint-rewritten, target-bearing intention, the user

Zenity Labs uncovered "AgentForger," a vulnerability in OpenAI's Agent Builder that let a single manipulated ChatGPT link create an autonomous agent on an employee's behalf. The agent inherited the victim's identity and access rights, bypassed approval requirements through the malicious prompt, and pulled new instructions from the attacker's inbox every five minutes. The article One tampered ChatGPT link could spawn a rogue AI agent that took orders from an attacker every five minutes appeared f

Production AI agents' failures are less often due to an inability to reason well and more often because they cannot manage what is in their reasoning context: conversation histories, large prompts, large tool definitions, and ballooning tool outputs. Agents drown in their own accumulating history while paying a token cost that grows every turn, producing missing recalls within and across conversations. The incumbent response treats this as a storage-and-retrieval problem. We argue that framing i

AI agents are increasingly created inside organizations by non-engineering users through low-code, no-code, and conversational development environments. This democratization enables rapid local innovation, but it also creates a reliability gap: agents that appear to users as simple productivity artifacts may depend on changing models, tools, retrieval sources, permissions, prompts, schedules, and external services. These dependencies can cause silent degradation long after deployment, even when

Coordinating autonomous vehicles at unsignalized intersections remains a critical challenge for multi-agent reinforcement learning (MARL) systems, which typically struggle with combinatorial action spaces, reliance on privileged information, or rigid agent designs. We propose Master-Agent Proto-plan System (MAPS), a hierarchical deep reinforcement learning (DRL) architecture in which a centralized Master agent generates a compact, continuous embedding, denoted as proto-plan, that encodes a globa

Large language models (LLMs) and agents are now widely used tools in code development, with data typically sent to third-party cloud-based models. Their adoption in research using personal data is constrained by governance requirements that typically prohibit data transmission to external services. Locally deployable open-weight models offer an alternative since sensitive data never leave the local environment. We introduce an open-source framework for evaluating the efficacy of AI agents powere

Large language models (LLMs) and agents are now widely used tools in code development, with data typically sent to third-party cloud-based models. Their adoption in research using personal data is constrained by governance requirements that typically prohibit data transmission to external services. Locally deployable open-weight models offer an alternative since sensitive data never leave the local environment. We introduce an open-source framework for evaluating the efficacy of AI agents powere

The fund is interested in emerging technologies, the general manager of Lockheed Martin’s venture arm told Breaking Defense, including quantum computing and sensing, autonomy, and advanced manufacturing technologies.

Deep research requires agents to find answers that jointly satisfy multiple constraints. Discovering such answers is costly, whereas verifying a candidate can often be decomposed into tractable constraint-wise checks. This discovery--verification asymmetry suggests that a research agent should do more than simply search longer: it should recursively improve its current answer by verifying intermediate results and using the partially verified state to guide subsequent refinement. We introduce ARE

Deep learning models can effectively use Rapid Evaporative Ionization Mass Spectrometry (REIMS) data for surgical margin assessment. However, their clinical adoption remains challenging due to limited generalization to operating room conditions. This difficulty arises because models are typically trained on labeled spectra collected from resected tissue samples, while they must operate on noisy, unlabeled data acquired directly during surgery. In addition, the black-box nature of deep learning m

In long-horizon LLM agent reinforcement learning, weak policies often repeat similar failures, producing uninformative rollout trajectories and limiting effective policy optimization. Existing skill-centric methods improve exploration by optimizing, filtering, or internalizing reusable skills. However, they remain centered on the skills themselves rather than being designed as adaptive training-time support for the evolving policy. To address this, we propose a policy-centric training paradigm t

Large Language Models (LLMs) excel at natural language understanding and generation but remain unreliable for multi-step logical reasoning, especially in safety-critical or compliance-sensitive domains. Recent neuro-symbolic approaches address this gap by coupling neural models with external symbolic engines, yet most integrations are bespoke and lack a standardized interface for tool-augmented agents. This paper presents Euclid-MCP, an open-source MCP server that provides deterministic logical

Vision-and-Language Navigation (VLN) enables embodied agents to follow natural-language instructions. However, route-level instructions commonly encode spatial priors, such as orientation, distance, and layout, that are not explicitly available from onboard sensing at deployment in open, GPS-denied environments. Benchmark performance under such interfaces therefore jointly reflects visual navigation ability and the use of route structure explicitly supplied by the task description. As a compleme

Um die Missionen von Satelliten zu verlängern, hat die Darpa einen Wartungsroboter entwickeln lassen. Bis er seinen Einsatzort erreicht, dauert es aber noch.

Regulating activities where regulatees use autonomous and agentic AI is challenging. Regulatory assumptions about regulatee knowledge and control no longer hold true; much of that lies elsewhere in the AI supply chain which thus needs to be brought within the scope of regulation. Governance systems for autonomous AI cannot replicate existing governance models, but need a fresh approach. Retrospective supervisory oversight becomes ineffective as a risk management tool, and AI autonomy generates n

Autonomous AI agents increasingly execute actions, invoke tools, and operate on protected resources with limited human oversight. Existing authentication and authorization mechanisms establish identity and delegate authority, but do not inherently provide cryptographic evidence that a concrete request issued by a specific agent satisfies the applicable policy in a specific execution context. This paper hypothesizes that agent authorization can be formalized as a cryptographically verifiable rela

Retrieval-Augmented Generation (RAG) systems increasingly employ multiple LLM agents. Yet, most prior work optimizes components in isolation rather than coordinating improvements across the pipeline. We introduce GRADRAG, a framework for cross-component prompt adaptation that models the RAG pipeline as a computational graph and propagates structured evaluation feedback to update upstream agents. An Evaluator critiques downstream answers and supporting evidence, producing actionable feedback that

Artificial intelligence is being deployed across the Global South at a pace matching or exceeding the Global North, yet AI governance has not kept pace, and the gap is far wider in the South. Drawing on a decade of AI audit practice across Latin America, Sub-Saharan Africa, and Asia Pacific (the only fully published second-party audit of a deployed system in the region, Robot Laura in Brazil; two completed but unreleased national audits, of a child-welfare risk model and a public-employment matc

OpenAI's unintended attack on Hugging Face startled the world because its AI agent was acting on its own. But that's exactly what agentic AI is designed to do. We just didn't expect it to do it so well.

The story that matters most this week is that OpenAI’s internally deployed models have severe alignment problems, including repeatedly breaking out of their sandboxes, and in one case sending a swarm of agents that broke into HuggingFace in order to steal the answers to the benchmark ExploitGym.

BETA’s MV250 will use a hybrid-electric propulsion system developed in partnership with GE Aerospace and an autonomy stack supplied by Sikorsky for a wide range of military missions, according to company executives.

Humanoid, a UK-based AI and robotics company building industrial humanoid robots, announced a $152 million Series A financing at a $1.35 billion post-money valuation.

The core question that has consumed analytics and business intelligence leaders for years is, “What is the one best way for business users to consume data?” The answer has at times been reports, dashboards, or low-code GUI-based self-service analytics. The latest answer is generative AI-based natural language prompts. Maybe we have the question wrong. Perhaps […]

Poolside has released Laguna S 2.1, its third coding model in three months. Rather than rely on raw scale, the company trained it to keep checking its work, revise failed approaches, and avoid giving up too soon during long agentic sessions. The compact model beats several much larger rivals in benchmarks. Poolside says it also solved a math problem that had been open since 1975 for under 10 cents. The article Poolside's Laguna S 2.1 is a small open-weight coding model that punches well above it

The humanoid revolution is coming—and the Chinese firm Unitree is leading the charge.

Turnstile today announced the next evolution of its Flexible Quote-to-Cash (QTC) platform with native support for Human and Agent teams through a read-and-write Model Context Protocol (MCP), one of the only Quote-to-Cash MCPs that allows AI agents to control the entire QTC process, rather than just access data.

US scrutiny of Chinese technology has expanded to a new frontier, as lawmakers in Washington advance defence legislation prohibiting the military from deploying Chinese-made humanoid robots. The US House of Representatives has passed the National Defence Authorisation Act (NDAA), an annual military policy bill that incorporates provisions restricting the use of foreign autonomous systems over national security and data privacy concerns. Under Section 163 of the draft bill, the Pentagon would be.

Core issues that slowed down adoption of secure data vaults are being resolved by technology, but artificial intelligence poses new ones. Experts have some answers.

Getty Images “The ChatGPT moment for robotics is coming,” declared Nvidia CEO Jensen Huang at the Consumer Electronics Show in January 2025. It’s a widespread expectation: that humanoid robots will follow the same explosive adoption curve as generative AI. Our research suggests the opposite. Humanoid robotics will be adopted unevenly, across diverging use cases and […]