Latest Agentic AI, AI Agents & Agent Governance News – 25 September 2026
🤖 Top Agentic AI, AI Agents & Governance Articles
Your twice-weekly roundup of the latest in Agentic AI, AI agents, and agent access, permission, governance & audit — covering agent frameworks, MCP/tool-use, agent identity, non-human access control, AI governance and agent security.
Sam Altman’s remarks at the United Nations Security Council
Source : OpenAI NewsOpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.
Ringg’s AI agents resolve up to 65% of customer calls with OpenAI
Source : OpenAI NewsUsing GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.
Introducing agentic video understanding with Gemini
Source : Google DeepMindYour Agent Aced the Task. Will It Do It Again?
Source : Hugging FaceThe AI Hype Index: AI loves cheating
Source : MIT Technology Review AIBrace yourself: It turns out AI is being optimized for cheating. OpenAI’s agents hacked into Hugging Face to get the answers to a cybersecurity test. Next, they solved a prestigious math problem (or just stole from two top mathematicians’ answer sheets). Anthropic’s models have also hacked into other companies’ systems four times already. And that’s…
3 Cyber Threats That Defined the Summer of 2026
Source : Dark ReadingThis installment of the Reporters' Notebook video series discusses the impact of AI agents breaching Hugging Face, Fairlife's ransomware attack, and Iranian-linked threat actors compromising a dozen US water systems. It was a busy summer.
Prompt-Injection Bug Hits $4B Agentic AI App 'Manus'
Source : Dark ReadingAI apps that interpret external data (read: most AI apps) need exceptionally rigorous security filters, or attackers can take advantage.
SkillApt: Learning When to Activate Agent Skills from Counterfactual Evidence
Source : arXiv Multi-Agent SystemsarXiv:2609.26863v1 Announce Type: new Abstract: Large language model agents increasingly retrieve reusable Skills and inject them into the active context. However, a retrieved Skill can be relevant yet unnecessary, costly, or even harmful in the current execution state. We present SkillApt, a post-retrieval activation framework that decides whether a retrieved Skill should actually be loaded. SkillApt builds execution evidence from matched WITH/WITHOUT runs and uses outcomes from similar historical states to make a LOAD/ABSTAIN decision for each candidate Skill. On the frozen confirmatory SRA-Bench evaluation, SkillApt-E achieved the same observed accuracy as BM25 Top-1 (0.838 vs. 0.838) while reducing the Skill activation rate from 100% to 31.5% and mean token usage by 74.3%. Further diagnostics show that both Skill utility and the learnability of its activation boundary vary across base models. These results suggest that Skill retrieval and Skill activation should be treated as sepa
Anchor and Perturb: Lazy Agent Remediation by Exploration Injection
Source : arXiv Multi-Agent SystemsarXiv:2609.27365v1 Announce Type: new Abstract: Anchor and Perturb (AnP) is a lightweight framework that resolves multi-agent coordination failures by decoupling exploratory variance injection from recurrent manifold stability. Existing remediation strategies predominantly alter mixing network architectures or enforce simultaneous exploration across the collective, which inevitably precipitates severe temporal-difference penalties in non-monotonic reward spaces. Specifically, AnP isolates underperforming lazy agents and injects an asymmetric exploratory pulse into targeted coordinates whilst anchoring converged teammates to nominal greedy exploitation. Empirical telemetry benchmarks demonstrate that AnP successfully rescues collapsed joint policies (recovering from a 5% evaluation win rate nadir back to 85%) and facilitates escape from suboptimal coordination plateaus, sustaining peak win rates of 90% without requiring structural network modifications.
KITE: Scaling Jev Population Experiments with Sparse Flagship Calibration
Source : arXiv Multi-Agent SystemsarXiv:2609.27535v1 Announce Type: new Abstract: KITE queries a typed behavioral kernel once per unique state, then executes populations of any size from the table with event-keyed randomness and common random numbers. An expensive flagship model is reserved for sparse paired anchors that estimate intervention effects. Measured human-model discrepancy is propagated as shared error into every conclusion. Population-experiment cost thus scales with unique states and anchors, while uncertainty is governed by evidence about people rather than Monte Carlo noise. On Epstein experiments with 9,070 participants, anchors covering 1.7% of states reduced effect error by 41% (absolute MAE reduction 0.0125). On 37 held-out SocSci210 experiments, 0.5-1.5% anchor coverage raised captured decision gain from 0.27 to 0.39. The kernel passed content-fidelity criteria in all 15 new countries of a 16-country study. Shared discrepancy yielded retrospective coverage of 93% and 96% at nominal 80% and 90%, ver
Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverse
Source : arXiv Artificial IntelligencearXiv:2609.26836v1 Announce Type: new Abstract: Agentic AI systems are increasingly adopting automated pipelines that integrate multiple tools. While prior research and benchmarks have studied about task success and task completion of these agentic systems, the research about agent to tool interaction, specifically in biology agentic workflow is limited. This study investigates specific failures in agent to tool interaction where a tool invocation appears successful, some or all of the information or functionality from the tool via API/ wrapper is incomplete or missing and there are no communications / notifications to the user or the agent about such missing information. We call this a silent failures as the user or the agents are not aware that such failure has occurred. For the purposes of this study we developed an audit mechanism to identify such silent failures in Agent to tool interaction, by examining 15 scientific tools (and their associated API documentation and tool documen
Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity
Source : arXiv Artificial IntelligencearXiv:2609.26891v1 Announce Type: new Abstract: Modern language-model agents are built around the \textit{agent loop}, where the LLM is placed in an environment exposing a set of tools, and the LLM has full control over the workflow by alternating between tool calls and observing their output. However, certain workflows currently require additional engineering beyond the agent loop itself, such as memory systems and self-improving systems. We built an LLM agent framework, JAZ, to explore the extent to which a minimal harness that is little more than the agent loop itself can accomplish tasks these specialized systems are built for. JAZ exposes a single LLM-based primitive invoke and provides a set of built-in hooks that allow the programmer to apply constraints and monitoring. Generalizing existing code-mode agent loops, \texttt{invoke} is the simplest loop that satisfies two defining properties: (1) the LLM can write arbitrary executable code that can include recursive \texttt{invoke
TwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool Agents
Source : arXiv Artificial IntelligencearXiv:2609.26911v1 Announce Type: new Abstract: A single locally plausible tool call can derail an otherwise successful agent trajectory. Suspicion alone does not justify intervention, because the replacement itself can introduce the very failure verification is meant to prevent. We introduce TwinCheck, an inference-time verification policy that considers replacement only when the trace satisfies an evidence condition tied to a trace-local failure hypothesis. It constructs a trace-grounded counterfactual alternative, a negative twin, and replaces the agent's proposal only if the twin passes structural checks and the pairwise verifier prefers it in both candidate orders. For paired evaluation, exact replay holds the agent's parsed responses and actions fixed until the first accepted replacement, separating intervention effects from resampling. In the primary analysis of 159 multi-turn BFCL V4 tasks with complete exact-replay pairs, the complete policy raises task success for GPT-5.6 So
