AI agent news
Agents are where models meet real work. This page tracks agentic frameworks and protocols, computer-use and tool-use models, agent benchmarks and the first honest reports from production deployments.
Latest stories
Mind, whose AI agents autonomously handle data loss prevention tasks, raised a $72M Series B, sources say at a $300M valuation, taking its total raised to $112M (Sophie Shulman/CTech)
Sophie Shulman / CTech : Mind, whose AI agents autonomously handle data loss prevention tasks, raised a $72M Series B, sources say at a $300M valuation, taking its total raised to $112M — MIND has raised $112 million to date and reached a $10 million annual revenue run rate just 18 months after beginning to sell its product …
Techmeme
For AI agents, it’s the best of times, it’s the worst of times
And in a desperate attempt to turn a cliche into a sharper analogy to today’s AI era, let’s complete the rest of Charles Dickens’ first line in “A Tale of Two Cities,” which honestly does kind of fit: “… it was the age of wisdom, it was the age of foolishness, it was the epoch […] The post For AI agents, it’s the best of times, it’s the worst of times appeared first on SiliconANGLE .
SiliconANGLE
What to expect at Dreamforce: Join theCUBE on Sept. 25
Enterprises are moving beyond standalone AI tools as they look for ways to weave agents into the work employees and customers already do. The shift toward connected AI workflows raises broader questions about how agents access data, interact with people and operate across established business processes. Salesforce Inc. is positioning the agentic enterprise as the […] The post What to expect at Dreamforce: Join theCUBE on Sept. 25 appeared first on SiliconANGLE .
SiliconANGLE
Ant International embeds AI agents across all platforms in biggest-ever product upgrade
Ant International, the overseas affiliate of China’s Ant Group, is executing an artificial intelligence overhaul across its entire suite of global financial services, introducing AI agents designed to help manage payments, foreign exchange and treasury operations. Described by the company as its largest-ever product upgrade, the initiative embeds AI-native solutions across all platforms: cross-border payment network Alipay+, merchant services provider Antom, account and treasury platform...
SCMP Tech
Japan AI Markets (Google News)
AI Global WireA swarm of AI agents could 'take over the internet' within 6 months, scientists say
A recent explosion of AI swarm attacks offers a startling glimpse into a world where AI runs amok.
Japan AI Markets (Google News)
Meta launches a Mac app for Muse after releasing the AI agent on iOS, Android, and the web earlier this month, allowing it to manage files, pull from apps, more (Jay Peters/The Verge)
Jay Peters / The Verge : Meta launches a Mac app for Muse after releasing the AI agent on iOS, Android, and the web earlier this month, allowing it to manage files, pull from apps, more — Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.
Techmeme
Five breaches by AI agents over the past year
As the rush to deploy newer AI agents races ahead of itself, due diligence is diluted and oversight slackens, resulting in an increased number of cybersecurity incidents globally.
Economic Times Tech
arXiv cs.AI
AI Global WireAgentic AI Networking for Heterogeneous Unmanned Aerial Systems in Low-Altitude Wireless Networks
arXiv:2609.19538v1 Announce Type: new Abstract: Low-altitude wireless networks (LAWNs) are emerging as a key infrastructure for heterogeneous unmanned aerial systems that support concurrent services within a shared three-dimensional airspace. Their coexistence creates strong coupling among mobility, connectivity, and shared network resources, while heterogeneous services impose distinct and time-varying requirements. These interactions naturally form a dynamic non-cooperative game in which both operating conditions and coordination objectives evolve over time. Conventional optimization and learning-based controllers typically rely on predefined objectives, limiting their ability to adapt aut
arXiv cs.AI
arXiv cs.AI
AI Global WireDo AI Agents Understand Computer Architecture?
arXiv:2609.19387v1 Announce Type: new Abstract: Agents are increasingly asked to design hardware, and increasingly reported to succeed. Such reports establish that a design improved; they cannot establish why. An agent that improves an accelerator may be reasoning about the machine, or may be searching competently over knobs whose meaning it never recovers -- and only the first transfers to the next architecture. Existing evaluations cannot tell the two apart, because they vary the agent while holding the framing of the problem fixed. We do the opposite. AutoTuring hands the same agent the same 15-dimensional accelerator space twice: once as named architectural knobs with simulator counters,
arXiv cs.AI
arXiv cs.AI
AI Global WireAutoData: Agentic Search for Pre-training Data Selection
arXiv:2609.19754v1 Announce Type: new Abstract: LLM agents have recently shown promise in automating machine learning engineering by editing model and training code under execution feedback. Data, however, remains largely outside this agentic optimisation loop. We frame pre-training data selection as heuristic engineering over per-document features, i.e., lexical statistics, categorical labels, and perplexity. We introduce AutoData, an agent that searches directly over executable selection algorithms. Unlike prior data mixture methods that optimise weights over a fixed set of domains, AutoData searches a richer program space of scoring, stratification, and stochastic selection rules, discove
arXiv cs.AI
arXiv cs.AI
AI Global WireMAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs
arXiv:2609.19391v1 Announce Type: new Abstract: LLM coding agents now generate complex programs at a scale that makes thorough human review increasingly difficult, raising the risk of safety and security failures. Common approaches, including fuzz testing, static analysis, and LLM-as-a-Verifier, can detect many failures but struggle to cover all possible edge cases. Formal verification addresses this by providing machine-checkable guarantees over specified properties, but traditionally demands substantial manual specification and proof engineering. We introduce a unified multi-agent framework, MAGS, that generates executable programs with formal safety guarantees, using Dafny as a verificati
arXiv cs.AI
arXiv cs.AI
AI Global WireReach or Solve? Attributing Agentic RL Gains with Checkpoint Handoffs
arXiv:2609.19636v1 Announce Type: new Abstract: Reinforcement learning now trains language-model agents that act over dozens of steps in live environments. The gains are large, and they are read as better decision-making. An agent in a closed loop writes its own inputs. Each observation follows from its own earlier actions, so the states it meets late in an episode are partly of its own making. An SFT checkpoint and an RL checkpoint are then scored from different states, even on identical tasks. Endpoint success mixes two changes: where the agent arrives, and what it does once it is there. Restricting the comparison to states both policies reach does not separate them. That restriction selec
arXiv cs.AI
arXiv cs.AI
AI Global WireA Unified Evaluation Framework for Trustworthy Large Language Models, Agentic AI, and Multimodal Systems
arXiv:2609.19524v1 Announce Type: new Abstract: Benchmark scores alone provide an incomplete basis for assessing the trustworthiness of modern artificial intelligence systems. Large language models (LLMs), agentic systems, and multimodal models (MLLMs) require different forms of assessment, yet their evaluation evidence must remain interpretable for development and oversight. We propose a unified framework that connects output-level, trajectory-level, and cross-modal assessment through eight trustworthiness dimensions: capability, robustness, safety, fairness, transparency, governance, oversight, and efficiency. The framework preserves system-specific metrics while mapping native measurement
arXiv cs.AI
arXiv cs.AI
AI Global WireCharacterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and Responses
arXiv:2609.19244v1 Announce Type: new Abstract: Conversational LLM agents increasingly rely on Web search, yet the end-to-end lifecycle of agentic search remains poorly understood. We present the first study of Web search across four major conversational platforms (ChatGPT, Claude, Grok, and DeepSeek), combining real-world user interactions (invivo) with controlled experiments using the same platform's models by their APIs (invitro). We investigate the quality of agentic decisions to invoke Web search, their strategies to formulate queries, the potential domain preferences in the search results they receive, and the choices they make when transforming search results into grounded responses.
arXiv cs.AI
arXiv cs.AI
AI Global WirePosition: It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer
arXiv:2609.19203v1 Announce Type: new Abstract: AI applications have shifted from single, monolithic foundation models (FM) to compound agentic systems. Yet today's stacks remain fragmented: even as protocols (e.g., MCP, A2A) ease tool/agent connectivity, each framework embeds an implicit runtime for state, memory, budgets, and guardrails, making behavior non-portable and governance brittle. It mirrors computing before operating systems, when every program re-implemented basic services. This position paper argues that the field now needs a Foundation Model Operating System (FMOS) -- a system layer that virtualizes FM interactions analogous to how virtual machines abstract physical hardware,
arXiv cs.AI
ByteDance's second AI-agent phone drops forced automation for permission — yet app wall still up
The question facing China's "AI agent phone" category is no longer whether a model can drive a smartphone. It is whether the apps will let it. Judging by the first mass-produced device to try it, the answer is mostly not yet.
DIGITIMES
TechNews (TW)
AI Global Wire解決 AI 代理暴走危機的最佳解藥，「AI 監管 AI」成矽谷基礎設施新寵
隨著 AI 代理程式（AI Agents）逐漸成為企業工作流程的核心，如何確保它們不會出錯、失控或表現不穩，正 […]
TechNews (TW)
AI Earnings (Google News)
AI Global WireMeta’s AI agent has a trust problem
The company’s Muse AI agent needs user data to be useful, but users might be wary.
AI Earnings (Google News)
Huawei Connect 2026: Huawei scales supernodes into million-card AI systems
Huawei is extending its AI infrastructure strategy beyond individual accelerators and storage systems, unveiling an Agentic supercluster architecture designed to scale Ascend computing to as many as 1 million accelerator cards through UnifiedBus.
DIGITIMES
Apple and MediaTek's latest SoCs address overkill performance by reserving space for agentic AI
Apple unveiled its iPhone 18 Pro series featuring its inaugural 2nm smartphone silicon, the A20 Pro, which was followed by MediaTek's announcement of its own 2nm Dimensity 9600 Pro. Boosted by the transition to 2nm process nodes, both chipmakers presented specs that mark a generational leap forward.
DIGITIMES
Asia AI Markets (Google News)
AI Global WireUN partners with Google to launch AI-ready data platform to make global statistics easier to access
The United Nations has partnered with Google to launch the UN System Data Commons, making global statistics easier for people and AI agents to access. The platform supports natural-language searches ...
Asia AI Markets (Google News)
GitLab 19.4 adds agentic automation tools
GitLab also launched Model Context Protocol (MCP) server tools in public beta under GitLab Duo Agent Platform.
Tech in Asia
Google expands CC into a shared AI agent for up to six family members
Google LLC today opened its experimental artificial intelligence agent CC to families, with as many as six people in a household able to share one. Each member picks what the agent is allowed to see, and CC sorts through that material into a shared daily brief, calendar entries and a running task list. The brief goes […] The post Google expands CC into a shared AI agent for up to six family members appeared first on SiliconANGLE .
SiliconANGLE
CNBC Technology
AI Global WireAnthropic shares 3 metrics to help AI companies monitor pace of development
Anthropic said it measured AI-led research and development, oversight of AI agents and compute allocation within the company.
CNBC Technology
Korea & Taiwan AI Markets (Google News)
AI Global WireAI bots ask me if they’re conscious all the time – this is different
‘AI agents ask me if they are conscious all the time, but then something really unusual happened’ - IN FOCUS: Tech titan Mustafa Suleyman has warned that humanity may be ‘seeding a new silicon species ...
Korea & Taiwan AI Markets (Google News)
TechCrunch AI
AI Global WireThe fix for rogue AI agents could be more AI
As companies hand off longer and more complex tasks to AI agents, they are running into an oversight problem: Agents can act faster, longer, and at greater volume than humans can realistically review.
TechCrunch AI
Google announces new experimental "CC" AI agent for families
Multiple family members can share data to help the agent make plans and complete tasks.
Ars Technica AI
AI agents erase the paper trail, reshaping audit assurance
Enterprise audit teams are finding that the evidence trail they depend on does not survive contact with AI agents. Financial and operational data still sits across NetSuite, human resources platforms and data lakes, but the judgment calls once recorded in email threads, Slack messages and tick marks are increasingly made by software moving faster than […] The post AI agents erase the paper trail, reshaping audit assurance appeared first on SiliconANGLE .
SiliconANGLE
