< ciso
brief />
Tag Banner

All news with #google tag

796 articles · page 3 of 40

BigQuery AI.AGG: Semantic Aggregation at Scale

🧭 BigQuery introduces AI.AGG, a preview multi-level semantic aggregation function that enables natural-language instructions over millions of rows using a single SQL query. The engine orchestrates row validation, context-safe batching, parallel inference with Vertex AI, and recursive synthesis to produce final group-level summaries. AI.AGG supports text and multimodal inputs, integrates with BigQuery Object Tables, and requires careful prompting and NULL handling for reliable production use.
read more →

Gemini Omni 1.2 Flash Now Generally Available

🔔 Gemini Omni 1.2 Flash is now generally available on Google Cloud, optimized for high-volume everyday video use cases like product videos, virtual try-ons, and ads. Built for speed and cost-effectiveness, it follows image, video, and gated audio references more closely and includes an imperceptible SynthID watermark. Business users can access it in Gemini for business and Google Vids, while developers can integrate via API and benefit from enterprise-grade infrastructure and Provisioned Throughput.
read more →

Google donates agent runtime stack to CNCF

🧭 Google is donating three open source projects—Agent Substrate, gVisor, and OpenRL—to the Cloud Native Computing Foundation, and Agent Substrate has reached 1.0 and is generally available on GKE. These projects provide an open stack for running large-scale, stateful autonomous agents across CPUs, GPUs, and TPUs by decoupling agent state from compute, sandboxing untrusted code, and orchestrating reinforcement learning post-training loops. The move aims to enable neutral governance, broader community contributions, and production-ready isolation and efficiency for agentic workloads.
read more →

AlloyDB Express: Instant, Agent‑Ready PostgreSQL

⚙️ AlloyDB Express launches in preview as a fast, agent-ready offering of AlloyDB for PostgreSQL aimed at developers and AI coding agents. Databases can be created in under two seconds, automatically pause after 15 minutes of inactivity, and resume in seconds, helping reduce cost and speed up testing and prototyping. Built-in enterprise features include vector and text search, direct access to Gemini models in SQL, secure networking, IAM controls, and a one-click upgrade path to production AlloyDB.
read more →

Google issues November Android security fixes

🛡️ Google released its November Android security bulletin addressing a Critical flaw, CVE-2026-58868, which can enable remote code execution without user interaction on Android 14–17 when protections are bypassed. The patch is delivered with security patch level 2026-11-01 or later; a broader set of fixes requires 2026-11-05. The bulletin lists 216 vulnerabilities, 23 rated Critical, and notes no confirmed exploitation as of early November.
read more →

Introducing Physical Engineering’s Last Exam Benchmark

🛠️ Today Google AI & Infrastructure and Mercor introduce the Physical Engineers' Last Exam (PLE), a new benchmark and open competition to test agents on mechanical, electrical, civil, and structural engineering tasks. PLE provides 29 public tasks, grading tools, and a leaderboard at ple-bench.com, and includes long-horizon, real-world project folders with diverse file types. The benchmark evaluates code compliance, safety, and cost efficiency, revealing a wide gap between frontier models and professional engineering requirements.
read more →

Google Warns of Rising LLMjacking Resource Hijacks

🛡️ Google warns that threat actors are increasingly hijacking enterprise access to Large Language Models (LLMs) to fraudulently consume costly compute and AI services. Attackers obtain API keys via phishing, leaked code, misconfigurations, or exposed tokens to run expensive workloads while victims incur the charges. Incidents span multiple regions and industries, with single-day fraudulent charges exceeding $100,000 and cumulative losses reaching millions in some cases.
read more →

Google scales defenses to combat APAC scam networks

🛡️ Google outlines expanded measures to counter rising scam losses across the Asia Pacific region, which reached up to $114.1 billion USD in 2025. The company is deploying built-in protections such as Fake Call Detection and Pixel scam alerts, expanding Android developer verification, and promoting RCS-based defenses. Google is also engaging in industry coordination through the Industry Accord and new alliances like JASA, while Google.org commits $6 million to local partners for community training, AI prototyping, and policy support.
read more →

Google warns of rising LLMJacking threat to enterprises

🛡️Google's Mandiant warns enterprises of a growing threat dubbed LLMJacking, where attackers bypass front-end chatbots to exploit backend large language model resources using exposed API keys and cloud tokens. Incidents this year span financial services, retail, technology and transportation, with fraudulent cloud charges exceeding $100,000 in a single day and up to $1m cumulatively before access was revoked. The ecosystem is maturing, with actors escalating privileges, abusing routing software, and establishing resale marketplaces that proxy and resell illicit LLM access.
read more →

Industry Coalition Disrupts Major Android TV Botnet

🔎 Google and industry partners dismantled a botnet that had enrolled over 1.3 million Android TV set-top boxes, which secretly rented users’ Internet connections to residential proxy services and enabled ad fraud. The GTIG report attributes the preinstalled malware to Videostrong Technology Co., and says the takedown should cut access to at least 10 million additional compromised IoT devices. Synthient and other partners corroborated findings that multiple proxy SDKs were used across affected devices.
read more →

Quick Google Security Checkup for Gmail Safety

🔒 Run a Google Security Checkup to see who has access to your Gmail inbox in about two minutes. Review signed-in devices, remove outdated third-party app access, and confirm your recovery information to secure your account. These three steps help reduce phishing risk and prevent unauthorized access. For additional guidance, consult the Gmail Help Center on protecting your account.
read more →

Agentic Offensive Frameworks Accelerate Intrusion Pace

🛡️ Since late summer 2026, Google Threat Intelligence Group (GTIG) and Mandiant have observed threat actors leveraging open-source agentic pentesting frameworks such as ARTEX, PentAGI, Hexstrike, and CAIRN to automate broad portions of the attack lifecycle. These multi-agent, LLM-driven tools rapidly chain reconnaissance, lateral movement, privilege escalation, and exfiltration steps, enabling intrusions at machine speed. GTIG and partners outline observed tradecraft, examples of exploitation, and recommend defensive best practices including control-plane isolation, short-lived credentials, least privilege, segmentation, and behavioral detection.
read more →

New Gemini courses for finance and legal teams

📘 Google Skills has released industry-specific video learning bundles to help financial services and legal professionals adopt Gemini and agentic AI. The short courses present practical use cases and step-by-step examples for real tasks, such as rapid SEC document review for advisors or automated contract compliance checks for legal teams. All videos are available at no cost, and a companion Gemini Agent Ready program offers credits and skill badges for advancing to production-grade agent development.
read more →

Google unveils Gemini 4 Carbon with cybersecurity lead

🛡️ Google announced Gemini 4 Carbon, a new frontier model focused on advanced cybersecurity and long-horizon capabilities, claiming leading scores on benchmarks like SEC-bench Pro, ExploitGym, and CWE-bench. The company highlighted Carbon's improvements in vulnerability detection, exploitation validation, and patching, and said it is more likely to seek human input under ambiguity and resist indirect prompt-injection attacks. The rollout mirrors Gemini 4 Argon, initially limited to select cyber defenders under the Fairwind Program while Google engages in U.S. pre-release testing.
read more →

Google opens one‑tap installs to registered app stores

🛡️ Google will permit approved third‑party Android app stores to perform one‑tap installs and background updates on devices running Android 17 QPR2 or later in supported countries, starting December 1. To qualify, stores must pass identity verification, submit their entire catalog and every new app/update to Google for malware scanning, and pay a $10,000 annual fee. Stores must still obtain explicit user consent for installs, show visible notifications for large downloads or background installs, and remain liable for their catalogs. The program excludes the US and allows IT admins to block registered stores on managed devices.
read more →

AlloyDB Omni Web UI Brings Cluster Observability

🛠️ The AlloyDB Omni Web UI is now generally available for Red Hat and container deployments, offering a read-only, production-ready console for cluster visibility. It exposes hardware telemetry, query performance, columnar engine diagnostics, and parameter auditing so administrators can troubleshoot without leaving local or enterprise environments. The interface supports profile management, integrates with Prometheus for system metrics, and surfaces AI and ecosystem shortcuts while operating strictly in read-only mode.
read more →

Using VPC Flow Logs to Troubleshoot GKE Workloads

📘 VPC Flow Logs in Google Cloud annotates sampled network flows with GKE metadata so ephemeral pod IPs are attributed to workloads, namespaces, and services. This gives workload-level L3/L4 visibility without extra agents and complements in-cluster observability. Flow Analyzer in Network Intelligence Center and upgraded Observability Analytics buckets let you query, group, and visualize traffic by cluster, namespace, workload, pod, service, and zone. The post includes walkthroughs using Online Boutique to illustrate common troubleshooting cases.
read more →

How Google Ships Agent Skills Like Products

🛠️ This post describes how Google’s Agent Skills repository evolved operational controls and automated gates to treat skills as living products. It explains six CI/CD gates — structural conformance, linters, semantic versioning, evaluation suite checks, link validation, and an AI review agent — and why each is necessary at scale. The article contrasts small-scale authoring with repository-level governance and describes concrete rules for metadata, scripts, and evaluation.
read more →

GKE Dataplane V2 for high-density agentic AI

🔒 This post explains how GKE Dataplane V2 uses eBPF-powered identity mapping, label filtering, and metadata migration to support sub-second provisioning and extreme pod density for agentic AI workloads. It highlights Agent Sandbox and Agent Substrate on GKE as solutions for secure, lightweight isolation and state rehydration at large scale. The article outlines techniques to avoid the 65,536 identity limit and to maintain low-latency startup for thousands of sandboxes.
read more →

Cloud SQL Blue‑Green Deployments for Minimal Downtime

🚀 Cloud SQL now offers fully managed blue-green deployments, starting with Cloud SQL for MySQL, to simplify major version upgrades and configuration changes with minimal downtime. By creating a continuously synchronized staging clone (green) of the live instance (blue), teams can test against real production data and perform a fast cutover while retaining the original instance as a fallback. The managed workflow automates cloning, replication, verification, and switchover, reducing upgrade risk and operational overhead.
read more →