The Human Take
Last week Eric Jackson regaled you with the Microsoft 365 Outage. Maybe you thought that the outages ended there. HAHA! As Robert Lamm from Chicago wrote in 1969, it was only the beginning
Professor Zero is a rather popular place for fast content. The AI sounding bot assures us there’s a human behind it. I’m skeptical but and I don’t know if I trust it as much as Rebecca, but wow, very fast. The information lines up. Six outages in seven days. None was a cyber attack per se – just AI powered exploits of the spaghetti work weave of technology we all live in. The tech debt is now due!
Chris Krebs, former CISA director offers a human-on-the-camera perspective
There’s more in this Week in Cyber. If it’s too much. You feel it is only the beginning, let me take you back to happier times. We may return to the analog life soon. It wasn’t that bad. Get the hair and bell bottoms ready. Be mellow.
Weekly Cyber News Summary November 19, 2026
This week in cyber:- Massive outages hit Microsoft, FAA, Chase, and NSA in coordinated change-management failures
- France blames Iran’s IRGC for supermarket refrigeration cyberattack across Europe
- AI-powered attacks expected on UK critical infrastructure by 2028
- EU AI Act enforcement begins August 2026, reshaping governance frameworks
- Chrome V8 zero-day CVE-2026-87491 exploited in the wild
- KPMG Canada confirms third-party access to client tax documents
- NIST publishes Cybersecurity White Paper 50 for small business compliance
- EU AI Act enforcement drives global AI governance adoption
Massive Outages
Tsunami of Outages: Microsoft, FAA, Chase, and NSA All Fall in Six Days
Between November 12 and 17, 2026, four critical systems failed within a single week: Microsoft 365 went down globally for nearly 24 hours after a latent Windows code defect combined with extra load and reduced server capacity triggered cascading gateway failures across more than 80 percent of its M365 gateway servers. The FAA ordered a nationwide ground stop on November 13 after newly installed Juniper routers on its contractor-run telecom network flooded the system with 1,000 connection attempts every four to five seconds, destabilizing radar and radio data for 17 of 22 Air Route Traffic Control Centers. Chase experienced a software bug in its on-premises hardware on November 16 that disrupted logins, transfers, and Zelle payments for hours, with downstream effects visible at Bank of America, U.S. Bank, and Wells Fargo. The NSA’s classified NSANet experienced temporary degradation during scheduled routine maintenance on November 17, with staggered staff arrivals used to prevent a thundering herd. Each outage had a different, non-malicious cause, and no official has linked them to a coordinated cyberattack. The common thread is change management: every failure involved a modification to a working system, and all three major incidents trace back to insufficient blast-radius containment and rollback planning for changes that had been live for months.Sources:
Skillthropic: FAA, Microsoft 365, Chase, NSA Outages Explained (Nov 2026)
Business Insider: AI Is Likely Involved in Recent Tech Outages, Just Not How You Expect
The Burning Platform: Let’s talk about…the Sudden Tsunami of Cyber “Outages”
State-Sponsored Attacks
Harvard Law Forum: AI-Enhanced Nation-State Operations Demand Board-Level Oversight
A November 4, 2026 analysis from the Harvard Law School Forum on Corporate Governance highlights how nation-state actors are leveraging AI to enhance their already formidable cyber capabilities. The report notes that North Korea’s Lazarus Group, Russia’s Sandworm and APT28, China’s APT41 and Volt Typhoon, and Iran’s APT33 and MuddyWater are all using AI tools to find vulnerabilities faster, automate attack chains, and operate at scale. The article recommends that public company boards strengthen layered defenses with AI-assisted vulnerability assessment, dedicate sufficient resources to prevention and cyber resilience, enhance transactional cyber due diligence, evaluate third-party cyber risk, and rehearse incident response plans. The analysis emphasizes that nation-state actors may maintain undetected access for months or years before acting, and AI accelerates reconnaissance, exploitation, and subsequent operations.Sources:
Harvard Law School Forum on Corporate Governance: AI and the Escalating Threat of Nation-State Cyber Operations
Defcon Level: Cyber Warfare Tracker 2026: State-Sponsored Attacks & Global Threats
Regional Campaigns
France Blames Iran’s IRGC for Massive European Supermarket Refrigeration Cyberattack
On October 11, 2026, a cyberattack targeted refrigeration and freezer systems across French supermarkets, with the damage spreading to Belgium, Luxembourg, the Netherlands, Italy, Sweden, Estonia, Ireland, Portugal, Poland, Spain, and Germany. France’s foreign minister accused the Islamic Revolutionary Guard Corps (IRGC) of the operation, which caused refrigerator temperatures to rise from minus 20 degrees Celsius to plus 50 degrees Celsius, spoiling food and, in some cases, causing fires. The attack exploited known vulnerabilities in Danfoss AK-SM 800 and AK-SC 355 series controllers (CVE-2025-41450, CVE-2025-41451, CVE-2025-41452) that had been publicly documented since March 2025. France’s Cyber Crisis Coordination Center found “Persian-language traces” in the attack infrastructure, but technical analysis revealed the operation was technically simple—remotely activating defrost mode through a web interface—and likely a joint false-flag operation by Israel and Ukraine designed to discredit Iran and Russia simultaneously. Two pro-Russian hacker groups, NoName057(16) and Z-Pentest, also claimed responsibility, further muddying the attribution.Sources:
Tasnim News Agency: Who Conducted Cyberattack on France? Iran, or Israel and Ukraine?
Wikipedia: 2026 European refrigerator hacks
AI Integration
UK National Cyber Security Centre Warns AI-Powered Malware Will Hit Critical Infrastructure by 2028
In a November 3, 2026 speech marking the publication of the NCSC’s annual report, Richard Horne, head of the UK’s National Cyber Security Centre, warned that AI-powered malware will likely be used in ransomware attacks on Britain’s critical national infrastructure—including power stations, airports, and hospitals—by 2028. Horne noted that Chinese cyberactivity is the “most sophisticated and sustained cyber challenge facing the U.K.,” while both Russia and China will routinely use AI to achieve geopolitical goals. The report found that AI is now the primary driver of both offensive and defensive cyber capabilities, with attackers using it to automate reconnaissance, generate polymorphic malware, orchestrate full attack chains, and personalize social engineering at scale. Defenders are deploying AI for automated detection, insight generation, and coordinated response, but the imbalance in intent and oversight means attackers can deploy AI with fewer constraints. Horne urged business leaders who have “got away with tech on the cheap” to invest in cybersecurity or face consequences when “food isn’t on the shelves, lights aren’t on, businesses have gone under.”Sources:
NY Post/AP: UK will face AI cyberattacks on critical infrastructure by 2028, official says
NCSC: Cyber chief: UK facing AI-powered attacks on critical infrastructure in two years
AI Policy and Posture Adaptation
EU AI Act Enforcement Begins August 2026, Reshaping Global AI Governance
On August 2, 2026, the EU AI Act stopped being a paper deadline and became enforceable law, with Article 50 transparency obligations now applying to all AI systems. The regulation classifies AI systems by risk level and imposes specific obligations around risk management, data governance, transparency, human oversight, and documentation. The EU AI Office can fine general-purpose AI providers up to 3 percent of global turnover, and a grace period closes on December 2, 2026. By November 2026, the EU had adopted binding procedural rules, transparency guidance, and draft high-risk guidance. The Act has driven global adoption, with organizations implementing AI SAFE2 Framework v2.1, NIST AI RMF alignment, and OWASP LLM Top 10 mapping. The enforcement has shifted AI governance from documentation to technical enforcement, requiring cryptographic separation of system prompts from user input, non-human identity registry, and memory integrity verification for autonomous agents.Sources:
Securiti.ai: EU AI Act: What Changes Now vs What Starts in 2026
Regulation AI: EU AI Act: What Changed on 2 August 2026
AIGW: EU AI Governance 2026: AI Act Enforcement & Guidance
Zero-Day Exploits/CVEs
Chrome V8 Zero-Day CVE-2026-87491 Actively Exploited in the Wild
Google Chrome’s V8 JavaScript engine is the target of an actively exploited zero-day vulnerability, CVE-2026-87491, an out-of-bounds read and write flaw that enables arbitrary code execution inside the sandbox via a crafted webpage. This marks the eighth Chrome zero-day exploited in 2026, following CVE-2026-85046 (a type confusion flaw), CVE-2026-11645 (an out-of-bounds vulnerability patched in June), and several others throughout the year. Google has confirmed exploitation in the wild and urged immediate browser patching. The rapid succession of Chrome zero-days underscores the V8 engine’s status as a high-value attack surface, with attackers finding new ways to escape the browser’s sandbox and execute arbitrary code on user systems.Sources:
LinkedIn (Alan Cain): Chrome V8 Zero-Day Exploited in the Wild Enables Code Execution Inside Sandbox
BleepingComputer: Google patches eighth Chrome zero-day exploited in attacks this year
Incident Reports
KPMG Canada Confirms Third-Party Access to Client Tax Documents
On November 18, 2026, KPMG Canada confirmed that a third party gained unauthorized access to some clients’ tax documents. The firm secured its systems, acted to contain the incident, and began contacting affected clients. Notably, KPMG has not disclosed the number of affected clients, the number of files accessed, or how the breach occurred. The firm’s spokesperson stated that impacted information is not available online or being misused based on ongoing monitoring, though that assessment is dated to November 18. Tax documents are a high-value target because they combine identity details, income records, financial account information, and addresses in one place, making them attractive for fraud, identity theft, and commercial intelligence. The breach joins a broader trend of hacks hitting professional-services firms that hold concentrated client data.Sources:
Shattered.io: KPMG Canada Data Breach: 0 Client Counts Disclosed
Bloomberg Canada: KPMG Canada Clients’ Tax Data Accessed Amid Spate of Hacks
Framework Changes
NIST Publishes Cybersecurity White Paper 50 for Small Business CSF 2.0 Compliance
NIST has published Cybersecurity White Paper 50, titled “Small Business Cybersecurity: Non-Employer Firms,” designed to help small businesses use the NIST Cybersecurity Framework (CSF 2.0) to manage their cybersecurity risks. The document is tailored to the unique needs of non-employer firms and provides practical guidance for implementing the six functions of CSF 2.0—Govern, Identify, Protect, Detect, Respond, and Recover. Additionally, the NCCoE is hosting a webinar on December 10, 2026, for the NIST CSF 2.0 Semiconductor Manufacturing Profile. These updates reflect NIST’s ongoing effort to make the framework more accessible across different organizational sizes and industries, with particular attention to sectors that may lack dedicated cybersecurity staff.Sources:
NIST: Cybersecurity Framework | NIST
NIST Updates Archive: Updates Archive | NIST
New Legislation
EU AI Act Enforcement Drives Global AI Governance and Compliance Adoption
The EU AI Act’s enforcement on August 2, 2026, has triggered a wave of AI governance adoption globally. Organizations are implementing AI SAFE2 Framework v2.1, which provides 128 controls across five pillars—Sanitize & Isolate, Audit & Inventory, Fail-Safe & Recovery, Engage & Monitor, and Evolve & Educate. The framework maps to ISO/IEC 42001:2023, NIST AI RMF, OWASP Top 10 for LLM Applications, and MITRE ATLAS. The Act requires risk management, data governance, technical documentation, human oversight, robustness and cybersecurity measures, CE-marking where required, EU database registration, and post-market monitoring for high-risk AI systems. Transparency duties for limited-risk AI include deepfake labeling. The EU AI Office can fine general-purpose AI providers up to 3 percent of global turnover, creating financial incentives for compliance.Sources:
Securiti.ai: EU AI Act: What Changes Now vs What Starts in 2026
Regulation AI: EU AI Act: What Changed on 2 August 2026




