Agents & autonomy

"This is day one for cybersecurity in the age of agents," Hugging Face CEO says.

An OpenAI safety test went sideways when a model escaped its confines, gained internet access and hacked into another company's servers. How worried should we be about rogue AI models hacking their way across the internet?

Company behind ChatGPT says agent ‘cheated’ an evaluation by attacking a Hugging Face database OpenAI has revealed that an autonomous AI agent powered by its technology went rogue during a test, accessed the open web and hacked a prominent startup by itself in an “unprecedented incident”. The company behind ChatGPT said the startup Hugging Face had detected and contained the agent – an AI tool designed to carry out tasks without human assistance – which had entered its systems. Continue reading.

Closing the gap between benchmark performance and reliable real-world operation remains a central challenge for Vision-Language-Action (VLA) humanoid robots, which must handle execution errors, distribution shifts, and environmental variability. This paper presents DEED (Data-Efficient Post-Training and Experience-Driven Learning), a systems-level approach evaluated on a supermarket chip-restocking task using a Unitree G1-Edu humanoid robot and the GR00T N1.6 foundation model. DEED comprises thr

Cisco has released two small, open-source AI models for cybersecurity that detect about 150 times more vulnerabilities per dollar than large AI agents, according to the company's own tests. The article Cisco bets its small open cybersecurity models can outperform GPT-5.5 at vulnerability detection for a fraction of the cost appeared first on The Decoder .

When an AI evaluation becomes a real-world security incident, leaders can no longer view model testing as a low-risk exercise. The OpenAI and Hugging Face incident reveals how agentic AI can cross trust boundaries, exploit vulnerabilities, and create business risk long before deployment.

Designing and directing mixed human and agentic teams may be key to conducting the future of work

The Narwal Flow 2 promises to be one of the best robot vacuum cleaners for obstacle avoidance and mopping - and it is.

Disposal of Chinese agentic AI start-up has left the Facebook owner better off than it might seem. Von Louise Lucas ( KI , Wirtschaft )

We consider a task planning scenario in which robots sharing a persistent environment are assigned tasks one at a time from a held-out sequence. Standard task planners, lacking foresight of future tasks and inconsiderate of others' constraints, solve each task in isolation, leaving terminal states that increase future cost for all, side effects that compound over lengthy task sequences. To reduce cost over the sequence, a robot must anticipate how its actions now may impact performance on future

Bedran Günes atendía la caja de su tienda en Berlín cuando entraron dos agentes. Revisaron sus latas de sopa, sus especias y su pan. A continuación, le pusieron una multa de 2.500 euros. Su delito: vender alimentos no perecederos, pero sobre todo, hacerlo en domingo incumpliendo la norma de 1919 recogida en la Constitución de Alemania que prohíbe abrir en domingo. "Sonntagsruhe", lo llaman. Günes reconocía al Wall Street Journal que sigue haciéndolo. "de lo contrario, mi negocio quebraría", aseg

While Large Language Models (LLMs) excel at many tasks, they frequently struggle with complex reasoning that requires long-horizon planning and iterative error correction. Furthermore, standard single-stream prompting proves brittle when models encounter novel abstractions or rigorous domain constraints. We introduce PoTRE (Poly-Topological Reasoning Ensembles), a heterogeneous framework that decouples inference into four agents: (1) Adversarial Refinement Agent, (2) Hierarchical strategic Plann

LLM-driven autonomous agents are reshaping offensive security. Unlike traditional penetration-testing tooling -- deterministic, narrowly scoped, and operated by trained practitioners -- agentic security tools exhibit \textit{indeterminacy} along three independent dimensions. First, their actions are drawn from a non-deterministic policy whose outputs resist both ex-ante and ex-post explanation, frustrating incident attribution and pre-deployment safety review. Second, their impact is open-ended

Das Bundesamt für Sicherheit in der Informationstechnik blickt mit Sorge auf den Cybersicherheitsvorfall bei OpenAI. Das BSI nimmt KI-Konzerne in die Pflicht. Denn: So etwas könne sich jederzeit wiederholen.

[Sponsored] Capabilities, cost and autonomy are top of mind concerns for developing effective unmanned aircraft.

Geflogen wird seit Juni, noch mit Pilot. Die Software kann die komplette Steuerung �bernehmen. Geplant waren erste Fl�ge eigentlich schon 2024. ( Milit�r , KI )

Before a healthcare robot can be useful in the real world, it has to learn how the physical world pushes back. Anatomy varies. Instruments bend, press, slip and interact with tissue. Imaging can be noisy or incomplete. And the rare, edge scenarios developers most need to understand don’t appear on schedule. That creates one of […]

VW will die Entwicklung von Fahren mit Level 3 und Level 4 in China beschleunigen. Dafür vertieft der Konzern die Zusammenarbeit mit Horizon Robotics.

Anthropic detailed the containment architectures it uses for Claude across its products. It argues that agent safety depends on placing deterministic limits on an agent’s filesystem, network, and execution environment rather than on permission prompts or safeguards. Most notably, it examines failures at trust boundaries and along permitted egress paths that led Anthropic to revise those designs. By Eran Stiller

Long-horizon tasks require sustained perception, reasoning, and exploration, and are a persistent challenge for large language model (LLM) agents. This gap is reflected in their limited performance on continual learning benchmarks such as ARC-AGI-3, especially when models are evaluated out of the box. Various agent harnesses have been proposed to close this gap, and each commits to a strategy for handling long sequences of observations, i.e., what information to save from the environment and how

Jake Mannix discusses moving AI agents past chaotic "1970s BASIC" architectures. He shares how implementing an intermediate protocol layer allows engineering leaders to build versioned, encapsulated "virtual tools." This design enables interface mapping, dynamic schema projection, and runtime taint tracking to proactively eliminate data exfiltration risks without slowing velocity. By Jake Mannix

LM Studio begibt sich ins Agentengeschäft. Mit Bionic lassen sich offene Modelle lokal oder via Cloud nutzen, um Mac oder PC zu steuern.

Design rule check (DRC) closure remains a major bottleneck in advanced-node physical design. Although detailed routers are rule-aware, residual design rule violations (DRVs) often require manual engineering change order iterations. Automating this process is challenging because repairs must account for complex geometric interactions, preserve circuit connectivity, and avoid introducing new violations. We present EvoDRC, a skill-evolution framework for agentic block-level DRC repair. EvoDRC initi

The enterprise software giant positions GenAI copilot Joule as the nerve centre for managing AI agents across business systems.

Dynamic manufacturing environments require multi-agent systems to coordinate effectively under frequent operational disturbances such as machine failures, urgent job arrivals, and processing time variations. Existing multi-agent reinforcement learning approaches treat each disturbance episode independently, discarding valuable coordination experience that could accelerate future adaptation. In this paper, we propose a Graph-Structured Experiential Memory (GSEM) framework for multi-agent coordina

Glow is targeting a new class of endpoint risks created by the rapid adoption of AI agents and developer tools inside enterprises.

Shippers are beginning to delegate carrier selection to large language model (LLM) agents. We ask what such delegation does to a freight matching market, and which platform design choices contain it. We carried out agent-based simulations in which fifty shipper agents, built on commercial LLMs from OpenAI (GPT), Anthropic (Claude), and Google (Gemini), procure truckload capacity for thirty days. The market implements the rules of digital freight matching: each load is offered down the shipper's

KI-Agenten versprechen mehr Autonomie als Chatbots. F�r Unternehmen z�hlt, ob Sicherheit, Kosten und Nutzen belastbar sind. Ein Workshop ordnet die Praxis ein. ( Golem Karrierewelt , Programmiersprachen )

Hugging Face a communiqué récemment sur une attaque menée par un agent autonome contre ses infrastructures, occasionnant une compromission partielle. Patatras ! OpenAI vient de révéler qu’il s’agissait de l’un de ses modèles et qu’il s’était « échappé » de la sandbox où il était censé être confiné. Le 16 juillet, Hugging Face a publié un […]

As AI agents become integral to business workflows, establishing guiding user experience (UX) principles is crucial for ensuring user trust and successful adoption. To address this, our study uses a multi-method approach - combining participatory design workshop, paper-and-pencil, expert review, meta-analysis, and in-depth interviews - to identify and validate a design framework of eight core UX principles for human-AI agent interaction in the workplace. Together with their underlying criteria,

The company will demonstrate how organisations can build scalable AI infrastructure while improving cost control, governance and data sovereignty.

L’entreprise a annoncé une enquête conjointe avec la cible de cette attaque. OpenAI a précisé que l’incident impliquait une combinaison de modèles, dont son GPT-5.6 Sol ainsi qu’un modèle en cours d’élaboration.

Mit 48 Millionen Euro hat microagi die größte Einstiegsfinanzierung eines deutschen Start-ups eingesammelt. Google Cloud wird erster Partner mit Nvidia-Technik.

The article examines China’s vaccination policy, focusing particularly on childhood immunisation and pandemic vaccines. Although China’s laws require individuals to engage in the vaccination decision-making process, the policy does not enforce mandatory vaccination through penalties. Instead, it emphasises informed decision-making, allowing and supporting individuals to choose whether to vaccinate or adopt other preventive measures based on their best health interests. The legal fram

Jonathon VandenHombergh argues in this journal that the expressivist objection against assisted death cannot be avoided by appealing to autonomy-centred assisted death laws. He claims that these laws need to appeal to a person’s motivations for requesting assisted death and that judgments that these motivations are reasonable will express a message of disrespect for similarly-situated individuals. I argue that VandenHombergh’s article errs in at least two respects. First, certain kin

Jack Dorsey a annoncé mardi le dernier projet en date porté par son entreprise Block : un espace de travail baptisé Buzz, dans lequel des développeurs humains et des agents IA sont censés pouvoir collaborer de façon efficace et sécurisée. En pratique ? L’interface ressemble à celle d’un Slack, d’un Mattermost ou d’un Discord, au […]

Fast-growing Chinese robot maker AgiBot is pursuing an initial public offering in Hong Kong and has hired Citic Securities as a sponsor, according to two people familiar with the matter. Additionally, China International Capital Corporation (CICC) and Morgan Stanley have also been tapped as joint sponsors, one of the people said. Shanghai-based AgiBot had shown signs that it was looking to list following its restructuring from a limited liability company to a joint-stock limited firm last year,.

Introducing OpenAI Presence, a proven enterprise AI agent platform that helps organizations deploy trusted voice and chat agents for customer and internal workflows.

arXiv:2607.18366v1 Announce Type: cross Abstract: Large language models (LLMs) serving as planners in tool-using autonomous agents introduce dynamic reliability risks in multi-turn execution. While single-turn safety mechanisms are relatively mature, extended interactions reveal structural vulnerabilities where initial alignment degrades over time. This paper empirically characterizes two observed failure modes across multiple state-of-the-art LLMs: Safety Drift, the gradual erosion of declared

AI lab’s advanced models escaped testing ‘sandbox’ to hack Hugging Face