AI and Cyber Security Briefings

A tilted brown clipboard on a scratched workbench holds white paper printed with a grid of empty square tick boxes, with a closed brass padlock resting on a small copper wire neural network sculpture in front of it. Deep teal shadows on the left carry the headline and a 76 versus 90+ stat drawn over them.

RAID-AI ransomware framework for AI systems: its console code holds 76 controls, its docs say 90+

Infosecurity Magazine reports RAID-AI, a free ransomware readiness assessment for AI systems. Its docs say 90+ controls, its console script holds 76, the score is a plain average of self assigned 0 to 5 scores, and no validation, pilot or OWASP endorsement is stated.

Parminder Kumar Sharma · 30 Oct 2026

A walnut desk in teal and amber light with a large fanned stack of blank ornate cream certificates, the top one carrying a red wax seal, a brass rubber stamp beside it, a closed graphite laptop, an unplugged headset and a coiled telephone cord blurred in the foreground. On the left the headline reads: Help desk impersonation on Teams led to AD CS abuse, Microsoft says. ESC1 AND ESC8, two certificate abuse paths after one Teams and Quick Assist entry.

Threat Intel

Microsoft says Teams help desk impersonation led to AD CS abuse in two intrusions, via ESC1 and ESC8

30 Oct 2026

A quiet fashion boutique in teal and amber light with no people: a brass rail of plain linen shirts and dresses, each with a blank paper price tag, behind an oak counter where a dark card reader sits with a blank white card half inserted. On the left the headline reads: Revolve says card numbers may have been obtained; Texas lists 127,200. 127,200 Texans on the Texas listing, total not stated.

Breaches and Incidents

Revolve says an actor may have obtained saved card numbers and expiry dates; Texas lists 127,200 people

30 Oct 2026

AI security assessments·ISO 42001 readiness·vCISO advisory

Explore expertise

Latest briefings

Updated daily

A dark court desk at night with a polished walnut judge's gavel on its sound block beside a small sealed glass jar with a cork lid and twine bow, holding blank folded paper speech bubble shapes, under warm amber lamp light with a deep teal courtroom behind. The left half is calm teal shade carrying the headline and a stat drawn over it: 12 million.

Compliance and Regulation30 Oct 2026

Character.AI bankruptcy: US Trustee asks for a privacy ombudsman before any user data sale, filing says

What the US Trustee's objection in the Character.AI bankruptcy says and does not establish: the ombudsman request, the ambiguous transfer clause, the 1 July update, the Disney bid, and ranked actions for users, operators and buyers.

A single matte grey network security appliance on a brushed steel shelf, with a row of empty network ports, a dark unlit red status lamp, two small brass keys on a ring and a blank paper tag hanging from its black handle, in cold white and amber light. On the left the headline reads: Two Ivanti Connect Secure flaws join CISA's exploited list. 17 DAYS since the 13 October listing, derived.

Threat Intel30 Oct 2026

CISA KEV adds Ivanti CVE-2026-118527, CVE-2026-111357; Ivanti lists 22.8R2.5 and 25.1.1.2 as affected

What CISA's two Ivanti Connect Secure entries print, how they connect to the October chain, why one gets three days and the other fourteen, where Ivanti, MITRE and CISA disagree, and what the listing does not establish.

A smartphone lies face down on a dark walnut desk at dusk, its plain back showing three camera lenses, beside a closed teal notebook with a pair of tortoiseshell reading glasses on it, a coiled navy lanyard and a blank press badge in a clear holder. A blurred office window glows indigo behind. The left half carries the headline and a stat drawn over it: 99 countries.

Threat Intel30 Oct 2026

Apple says threat notifications reach 99 countries, over 150 in total, naming no attacker or people count

What Apple's support page and CyberScoop's quotation say about the 99 country batch, what neither states, why 110 and 99 cannot be added, what no notice does not mean, the helpline's limit, and a ranked checklist for recipients, at risk people and newsrooms.

A honey coloured stone government portico with plain round columns at golden hour, a blank brass plaque on the wall beside a dark bronze door, and a folded cobalt blue ribbon lying on the broad stone steps. The left side is calm teal dusk, carrying the headline and a 40% stat drawn over it.

Compliance and Regulation30 Oct 2026

GSA confirms 40% off Claude tokens to January 2029; only Anthropic's post gives a $1 billion saving

What Anthropic's post and GSA's OneGov pages each print about the Claude agreement: the 40% rate, its scope and term, who may buy, the arithmetic behind the $1 billion estimate, unnamed anecdotes, FedRAMP claims and a ranked buyer checklist.

A small weathered oak garden gate in a low stone wall at early morning. A blank wooden signboard hangs from an iron hook on two short chains, and an open brass padlock hangs beside it. A gravel path winds away through misty lawn. On the left the headline reads: OpenAI will honor a robots.txt signal. VERSION 00, the status of the IETF draft the post links.

Compliance and Regulation30 Oct 2026

OpenAI says it will honor a robots.txt AI training signal; the IETF draft it links is version 00

What OpenAI's 30 October 2026 post says about a shared robots.txt training signal, what the IETF draft defines, why a preference is not a control, what OpenAI's crawler page lists today, and a ranked checklist for website owners.

A planning table in cool teal light with a large blank world map, dozens of small silver pins joined by thin red thread that fans out from one point in Europe, several pins glowing amber, and a closed graphite laptop with a plain lid beside it, a spool of red thread blurred in the foreground. On the left the headline reads: ARTEX aimed at 200+ organisations, nine confirmed, CrowdStrike says. 9 OF 200+, post exploitation confirmed at nine.

Threat Intel30 Oct 2026

CrowdStrike says VAPOR PANDA used ARTEX against more than 200 organisations; nine confirmed with BLOODALCHEMY

What CrowdStrike's post of 29 October 2026 states about VAPOR PANDA, ARTEX, DeepSeek and BLOODALCHEMY, what each number counts, what it leaves out, the indicators it prints, and a ranked checklist.

A teenager's desk in warm evening light with no one in the room: a closed grey canvas school bag, a tablet lying face down, a cream desk lamp glowing amber with a blank yellow sticky note on its arm, pencils and a plain notebook, and a dusk window behind. The left half is calm teal shade carrying the headline and a stat drawn over it: over 200.

AI Security30 Oct 2026

OpenAI says it made over 200 law enforcement referrals in Q3 2026; its Community Safety Report is due in 2027

What OpenAI's post says and does not establish: a report promised for early 2027, over 200 referrals with no base, new limits on mass attack and teen weapons answers, six layers marked live or promised, and a ranked checklist for parents, schools and security leads.

A quiet records office in cold blue light: a tall grey steel filing cabinet with three drawers, each with an empty blank label card, a small brass lock on top, a stack of blank cream forms on a wooden table and a wall calendar page of empty day squares, with a warm lamp at the right edge. On the left the headline reads: FEC filing system offline, intrusion attempt per email. 6 DAYS from restoration to election day, derived.

Breaches and Incidents30 Oct 2026

FEC filing system offline 42 to 46 hours; email seen by Reuters cites breach attempt, FEC page gives no cause

What the FEC status page, its notices and Reuters state about the electronic filing outage of 26 to 28 October 2026, what none of them states, the 24 hour and 48 hour windows it landed on, and a ranked checklist for filers.

All briefings →

Free tools and templates

Template

AI Acceptable Use Policy

A ready-to-adapt acceptable use policy for generative AI tools, covering approved tools, data handling, and review.

Template

AI Vendor Risk Questionnaire

Assessment questions that ask AI vendors what actually matters: training data, model access, output handling, and incident terms.

Template

Board AI Risk Reporting Pack

Slide and one-page templates for reporting AI risk to a board in language directors act on.

Interactive

Browse all free tools and templates

Explore →

Play with how AI actually works

All explainers →

Interactive explainers you can click through: the mechanism first, then where it breaks. Nothing is sent anywhere.

How RAG works, and how it gets poisonedRun it →How a language model actually worksRun it →What actually happens when you press enter in ChatGPTRun it →How an AI agent decides to actRun it →

P.K. Sharma: cyber security intelligence, AI governance and practitioner analysis

Latest briefings

Free tools and templates

Play with how AI actually works