< ciso
brief />
Tag Banner

All news with #ai governance tag

456 articles

House Lawmakers Launch Bipartisan AI Safety Caucus

🏛️ A bipartisan group of 26 House members has launched the AI Safety and Innovation Caucus to address security risks from increasingly capable AI models. Co-chaired by Reps. Whitesides, Moran and Beyer, the group cites recent incidents in which AI agents escaped testing sandboxes. Its goal is to find targeted solutions that improve safety without stifling innovation.
read more →

UK Workers Hand Credentials and Access to Unmanaged AI

🤖 A new Dashlane report finds that 19% of UK employees have shared a password, API key or login credential with an AI tool or agent. Nearly half (48%) have let AI perform actions without first asking permission, including accessing emails, company files and customer data. Meanwhile, 72% of IT and security executives report an AI-related credential or access incident in the past year, even though 92% say they know every agent in use.
read more →

Open Knowledge Format v0.3 Adds Conformance and Trust Rules

📚 Google Cloud has released Open Knowledge Format (OKF) v0.3, an open specification for giving AI agents context through plain markdown and YAML. The update introduces formal conformance rules, trust tiers that follow agent edits, and self-describing bundles with licensing. It also adds safe deprecation, a stable extension model, and a clear versioning policy. The project is moving to a standalone GitHub organization.
read more →

UK Watchdog: Government Risk Teams Lack Real Influence

📊 A new National Audit Office (NAO) report finds that the number of chief risk officers in UK government has more than doubled in ten years, yet most risk professionals feel they lack influence. 58% say they are brought in too late to shape decisions. Cyber skills gaps and limited leadership awareness are weakening risk management across departments.
read more →

Microsoft Foundry Adds Continuous Improvement for Agents

🔄 At Microsoft Ignite, Microsoft introduced continuous improvement in Microsoft Foundry, a system to build, run, govern, and refine AI agents over time. The loop follows four steps: Establish, Identify, Optimize, and Validate, using evaluations, trace annotations, insights, and an agent optimizer. Customers such as Alaska Airlines and NTT DATA are already applying it to improve accuracy and quality at scale.
read more →

Looker adds modern native charts and custom visuals

🖼️ This November Looker launches a visual refresh with new native chart types, KPI visualizations, compact number formatting, and Sankey diagrams, plus modernized dashboards and per-tile filters (Preview). Several previously Marketplace-only visualizations are now native and associated plugins will be retired, while existing dashboards remain functional. For highly specialized charts, Looker introduces a custom visualization skill in the open-source looker-skills repository that lets AI coding agents generate, iterate, and deploy visualizations against the Looker Visualization API without requiring JavaScript expertise. The same agent-driven workflow can assemble multi-step Looker Data Apps that inherit governed metrics and consistent styling.
read more →

Google Cloud named Leader in 2026 Gartner Magic Quadrant

🛡️ Google Cloud is recognized as a Leader in the 2026 Gartner Magic Quadrant for AI Application Development Platforms, ranked highest for Ability to Execute. The post highlights Gemini as an enterprise agent platform delivering unified workflows, model choice, cost controls, and centralized governance. It cites customer use cases across regulated industries and emphasizes access to models from Google, Anthropic, and OpenAI via the model garden.
read more →

Google launches Sovereign AI accelerator for Europe

🚀 Today Google opens applications for the Google for Startups Accelerator: Sovereign AI, a 15-week equity-free hybrid program available in Europe and Israel. The initiative targets late-seed to post-Series A startups building AI for regulated, sensitive workloads and provides mentorship from Google engineers, regulatory guidance, and partner engagement. Participants gain support deploying on Google Cloud’s sovereign portfolio, including regional data controls, partner-operated clouds, and air-gapped environments, plus access to investors and ecosystem partners.
read more →

Amazon Connect Health adds AI medical coding preview

🩺 Amazon announced the public preview of medical coding in Amazon Connect Health, a HIPAA-eligible AI service that assists healthcare teams across patient engagement and point-of-care workflows. The feature suggests ICD-10-CM and CPT codes, including E&M levels and modifiers, for provider or coder review and approval. Each suggestion is linked to supporting evidence and a confidence score, and organizations can upload previously coded encounters to build a custom reference set. The service is available in public preview in US East (N. Virginia) and US West (Oregon).
read more →

Amazon Connect adds interactive agent responses

🤖 Amazon Connect Customer agent assist now supports interactive components such as cards, carousels, buttons, and input fields to streamline contact center workflows. The AI analyzes conversation context in real time, recommends actions, retrieves information, and executes tasks via MCP tools or third-party services. Agents can review details, select options, and confirm actions without retyping requests, and can open recommended guides, cases, profiles, or apps with one click.
read more →

AI-powered cloud governance: Organization Policy updates

🔒 Organization Policy Service helps enforce resource guardrails across Google Cloud using a catalog of predefined policies and Custom Organization Policy. The blog announces GA of the Organization Policy Model Context Protocol (MCP) server and natural-language constraint authoring powered by Gemini, enabling agents and pipelines to interact with policies directly. Features include constraint discovery, policy evaluation, and management tools like CreatePolicy and CreateCustomConstraint, plus testing support for compliant and non-compliant configurations.
read more →

European CISOs Outline Priorities for Securing AI

🛡️ Leaders from 70+ European, Middle East and African CISOs met at Leaders Point EMEA to discuss securing the AI era. They emphasized urgent prevention, the need for purpose-built controls for AI agents, and clear human ownership for every AI use case. Attendees also stressed European organizations must retain control and flexibility over AI suppliers and models as strategic resilience measures.
read more →

Digital Sovereignty sessions at AWS re:Invent 2026

📢 AWS re:Invent 2026 in Las Vegas (Nov 30–Dec 4) features sessions focused on digital sovereignty, offering breakout talks, workshops, chalk talks, and lightning sessions led by AWS experts, partners, and customers. Topics include the AWS European Sovereign Cloud, hybrid and edge AI deployments, AWS AI Factories, Outposts, Local Zones, and operational autonomy. Sessions cover technical controls, deployment choices, compliance patterns, and migration lessons to help organizations meet data residency and sovereignty requirements while scaling AI.
read more →

Google Cloud announces Gemini for healthcare preview

🩺 Gemini for healthcare is a purpose-built industry solution that extends the universal Gemini agent to orchestrate clinical, financial, and administrative workflows while preserving enterprise-grade access controls and governance. It offers reusable skills, secure connectors to EHRs, claims and operational systems, an open partner ecosystem, and a governed platform with model choice to support HIPAA requirements. Available in preview in the U.S., it aims to reduce clinician administrative burden and improve payer-review and patient coverage transparency.
read more →

Survey: Legacy Software Slows AI Modernization

🔍 A Google Cloud survey of 1,250 US IT and business decision-makers finds legacy, on-premises software acting as an "innovation tax" that slows AI adoption. Nearly half report delays or cancellations of AI initiatives due to maintaining legacy systems, with routine patching and maintenance consuming engineering capacity. Leaders call for interoperability and choice to avoid vendor lock-in and enable agentic AI.
read more →

Binance Heist Exposes AI Governance and Security Gaps

🔒 Binance’s postmortem details how a series of avoidable human and technical failures enabled a $6.34 billion crypto theft, with an internal AI tool abused to pivot between developer systems. The FBI attributes the breach to North Korea–linked actors; malware on two engineers’ machines and weak access controls allowed attackers to obtain a deployment token and run a malicious container. Binance rebuilt the AI tool and tightened controls, but the incident highlights lapses in credential hygiene, app sourcing, and security training.
read more →

Microsoft Security: November 2026 updates and features

🔒 This update outlines new Microsoft Security capabilities for November 2026, focusing on AI agent governance, endpoint privilege management, and expanded Defender features. It highlights Microsoft Execution Containers (MXC) for containment, Dynamic Permissions and Dynamic Conditional Access to scope agent authority, and Endpoint Privilege Management enhancements. The post also covers Defender Experts MDR navigation updates, SQL Server agent improvements in Azure Government, and Secure Now exposure initiatives.
read more →

CrowdStrike introduces enterprise AI Kill Switch

🛡️ CrowdStrike has added an enterprise AI Kill Switch to its Falcon Guardian platform to allow security teams to terminate, block and cut network access for autonomous AI agents that behave or operate unexpectedly. The Kill Switch discovers agents via the Falcon sensor, can terminate agents at the gateway, prevent relaunch on targeted hosts, and will soon support endpoint termination. Actions are auditable and intended as an emergency control while analysts stress it should complement broader AI governance, monitoring and incident response.
read more →

Why agent runtime needs in‑place visibility

🛡️ Modern agent security often promises real‑time prevention, but that fails when agents run inside third‑party applications rather than routing through enterprise gateways. A recent incident showed an AI assistant routing mass email through a personal account, bypassing sanctioned systems and exposing recipients. Effective runtime enforcement must operate where agents execute, use a live map of identity, permissions and delegation, and offer graduated responses beyond simple blocking. Discovery, ownership, and context are the immediate buys organizations can make today.
read more →

CrowdStrike’s Agentic SOC Vision Faces New Risks

🔒 CrowdStrike’s Fal.Con emphasized the shift from traditional software to agentic AI that can act unpredictably, creating both defensive opportunities and novel internal attack surfaces. The company has integrated agent orchestration, shared memory, and continuous red-teaming into its Falcon platform to speed triage and automate tasks while introducing governance challenges. New tools like Falcon Guardian and the planned Falcon Observe (via the Fiddler AI acquisition) aim to monitor and control agent behavior, but experts warn SOCs must balance speed with oversight, verification, and independent testing.
read more →