{"id":5697,"date":"2026-08-17T18:00:36","date_gmt":"2026-08-17T23:00:36","guid":{"rendered":"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=5697"},"modified":"2026-08-17T18:00:36","modified_gmt":"2026-08-17T23:00:36","slug":"the-ciso-brief-august-16-2026","status":"publish","type":"post","link":"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=5697","title":{"rendered":"The CISO Brief \u2014 August 16, 2026"},"content":{"rendered":"<style>\n.single .entry-title,\n.single .entry-header .entry-title,\n.single .post-title,\n.single header.entry-header h1,\n.single h1.entry-title,\n.single .page-title,\n.post-template-default h1.entry-title,\n.post-template-default .entry-header,\narticle .entry-header,\narticle .entry-title { display: none !important; }\n.single .entry-header { margin: 0 !important; padding: 0 !important; }\n.single .entry-content { margin-top: 0 !important; padding-top: 0 !important; }\n<\/style>\n<table role=\"presentation\" class=\"wrapper\" cellpadding=\"0\" cellspacing=\"0\" border=\"0\" width=\"100%\">\n<tr>\n<td align=\"center\">\n<table role=\"presentation\" class=\"container\" cellpadding=\"0\" cellspacing=\"0\" border=\"0\" width=\"680\">\n<p>        <!-- Banner --><\/p>\n<tr>\n<td class=\"banner\" style=\"background-color:#0f172a;background:linear-gradient(135deg,#0f172a 0%,#1e3a8a 55%,#2563eb 100%);padding:36px 32px;color:#ffffff;\">\n<p class=\"date\" style=\"color:#ffffff !important;\">August 16, 2026 &middot; Weekly Edition<\/p>\n<h1 style=\"color:#ffffff !important;\">The CISO Brief<\/h1>\n<p class=\"tagline\" style=\"color:#ffffff !important;\">A presidential memo hands vetted private companies the authority to hack transnational criminal organisations &mdash; the biggest structural change to who is allowed to conduct offensive cyber operations in a generation. Around it, AI regulation stops being one debate and becomes four running in parallel: an FTC bid to police ideological bias, a GAO argument for amending laws we already have, Brussels putting seventeen Cyber Resilience Act standards out for comment, and California building its own programme. Underneath the policy noise, the plumbing gets attention too &mdash; Google Cloud dates its post-quantum finish line, and NIST asks how to rebuild the vulnerability database for machine-scale security data.<\/p>\n<\/td>\n<\/tr>\n<p>        <!-- At a glance --><\/p>\n<tr>\n<td class=\"content\">\n<h2>This week at a glance<\/h2>\n<p>The defining story of the week is a change in who is permitted to attack. On August 13 the White House issued a memorandum directing DOJ and DHS to stand up a programme under which vetted private companies may conduct surveillance and sabotage operations against foreign criminal organisations &mdash; bounded by rules against killing or injuring people, against anything amounting to a use of force under international law, and by a $1 million bond that is forfeit if a company breaks them. Cybersecurity Dive and CyberScoop both reported it, and both found the expert community genuinely split: supporters framing it as finally putting private capability behind public authority, critics reaching for the phrase &ldquo;letters of marque&rdquo; and asking what happens when a company misattributes a target and a foreign government treats the mistake as an act of state. Nextgov\/FCW&#8217;s companion essay on the dawn of the digital mercenary is the honest framing of what this is: warfare becoming an engineering procurement problem, with the same firms that sell drones and satellite connectivity to combatant commands now eligible to run cyber operations. For a CISO the memo is not abstract &mdash; if your organisation is in the security-vendor business, a new licensing question just landed, and if it isn&#8217;t, the escalation risk in your threat model just moved.<\/p>\n<p>AI regulation, meanwhile, stopped being one argument and became several at once. The FTC floated a policy statement treating ideological bias in AI systems as an unfair and deceptive practice under Section 5 &mdash; an unusual use of consumer-protection authority that would, by design, pre-empt state regimes like the Colorado AI Act. At the GovCIO AI Forum, GAO&#8217;s Nick Marinos made the opposite argument for a different problem: don&#8217;t write new AI statutes, amend the ones on the books, starting with the 2015 information-sharing law now running on a temporary extension through December 11. The White House signalled it may extend its voluntary safety-review framework to open-weight models once they reach frontier capability, which puts Meta, Nvidia and the open-source camp on notice. ETSI opened seventeen draft Cyber Resilience Act standards for comment ahead of the CRA&#8217;s December 2027 compliance date. California&#8217;s governor launched a state AI cybersecurity initiative that will require an AI cybersecurity officer in every state agency, explicitly citing insufficient federal support. And Transformer&#8217;s profile of Humans First &mdash; a conservative anti-AI grassroots operation incubated with Center for AI Safety money and run by a January 6 rally organiser &mdash; is a reminder that the politics underneath all of this are less stable than the policy documents suggest.<\/p>\n<p>The quieter engineering stories may matter more to next year&#8217;s budget. Google Cloud published a dated post-quantum roadmap: store-now-decrypt-later exposure mitigated across customer-facing workloads by the end of 2027, signature and identity protections complete by the end of 2028, general readiness in 2029, with work continuing into the 2030s. NIST issued a request for information on overhauling the National Vulnerability Database for an era of machine-scale disclosure and AI-assisted exploitation, conceding that periodic scanning, static prioritisation and manual remediation no longer scale. Both are the sort of multi-year programme that only gets funded if a CISO starts the conversation now.<\/p>\n<p>The rest of the week was the governance conversation catching up with itself. Dark Reading made the case that boards systematically underprice technology risk because deferred modernisation and technical debt never show up in quarterly earnings. CSO Online argued in one piece that the security function&#8217;s next move is helping the business say yes, and in another that a large vulnerability backlog is an organisational accountability failure rather than a security-team failure. SC Media&#8217;s contrarian take &mdash; that the talent shortage was never real, only a throughput problem &mdash; lands in the same conversation from the other end. Deloitte found just 16% of leaders consider their processes ready for agentic AI. And enforcement kept its own steady rhythm: the UK ICO reprimanded the Criminal Records Office over a breach that exposed data on more than 10,920 people, Make UK found only 51% of British manufacturers have a formal incident response plan, NCC Group pressed the case that a 1990 computer-misuse law still criminalises good-faith research, and the US judiciary agreed to start publishing how often judges authorise government hacking tools &mdash; in 2029.<\/p>\n<p>            <!-- Topic map --><\/p>\n<div class=\"topic-map\">\n              <img decoding=\"async\" src=\"https:\/\/www.cybersecurityinstitute.com\/blog\/wp-content\/uploads\/2026\/08\/topic-map-ciso-2026-08-16.png\" alt=\"Topic map of this week's CISO Brief themes\" loading=\"eager\"><\/p>\n<p class=\"caption\">This week&#8217;s topic map &mdash; the August 13 presidential memo and its DOJ\/DHS approval programme at the centre of a private-sector offensive-cyber cluster (letters of marque, the digital-mercenary thesis, Cyber Command deconfliction, transnational criminal organisations); a dense AI-regulation cluster spanning the FTC, GAO and the 2015 information-sharing law, open-weight safety reviews and OpenAI&#8217;s Astra pause, the EU AI Act and Cyber Resilience Act with ETSI&#8217;s harmonised standards, California, and the Humans First political backlash; the board-facing cluster of technical debt, the security backlog, the talent-shortage argument and Deloitte&#8217;s agentic-AI readiness gap; and the engineering cluster of post-quantum migration at Google Cloud and NIST&#8217;s NVD overhaul, connected through to legal exposure for security researchers, ICO enforcement and the Make UK manufacturing survey.<\/p>\n<p>              <!-- INTERACTIVE_MAP_LINK_START --><\/p>\n<p style=\"margin:10px 0 0;text-align:center;\"><a href=\"https:\/\/www.cybersecurityinstitute.com\/blog\/?p=5696\" target=\"_blank\" rel=\"noopener\" style=\"display:inline-block;padding:8px 18px;background-color:#0f172a;color:#ffffff !important;text-decoration:none;border-radius:6px;font-size:13px;font-weight:600;\">View interactive topic map &rarr;<\/a><\/p>\n<p><!-- INTERACTIVE_MAP_LINK_END -->\n            <\/div>\n<p>            <!-- Article index --><\/p>\n<h2>Article index<\/h2>\n<h3>Weekly News<\/h3>\n<h4>Washington authorises private-sector offensive cyber<\/h4>\n<div class=\"cluster-intro\">An August 13 presidential memorandum lets vetted companies conduct surveillance and sabotage against foreign criminal organisations, under DOJ\/DHS approval and a $1 million bond. Experts split sharply, and the &ldquo;letters of marque&rdquo; comparison arrives within hours.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>1. <a href=\"https:\/\/www.cybersecuritydive.com\/news\/us-private-companies-gangs-cyberattacks-offensive-operations\/827805\/\">US government will let private companies hack criminal gangs<\/a><\/td>\n<td class=\"src\">Cybersecurity Dive<\/td>\n<td class=\"dt\">Aug 13, 2026<\/td>\n<\/tr>\n<tr>\n<td>2. <a href=\"https:\/\/cyberscoop.com\/private-sector-hacking-presidential-memo-cybersecurity\/\">A bold new strategy or a dangerous precedent? Experts are divided on Trump&#8217;s memo<\/a><\/td>\n<td class=\"src\">CyberScoop<\/td>\n<td class=\"dt\">Aug 13, 2026<\/td>\n<\/tr>\n<tr>\n<td>3. <a href=\"https:\/\/www.nextgov.com\/cybersecurity\/2026\/08\/warfare-becomes-engineering-era-digital-mercenary-dawns\/415442\/\">As warfare becomes engineering, the era of the digital mercenary dawns<\/a><\/td>\n<td class=\"src\">Nextgov\/FCW<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>AI regulation moves on several fronts at once<\/h4>\n<div class=\"cluster-intro\">Four regulators, three jurisdictions, one week: the FTC reaching for Section 5, GAO arguing for amendment over new statute, the White House weighing safety reviews for open-weight models, ETSI opening seventeen CRA standards for comment, and California building its own programme rather than waiting.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>4. <a href=\"https:\/\/cyberscoop.com\/ftc-regulating-ai-ideological-bias\/\">The FTC wants to regulate AI for ideological bias<\/a><\/td>\n<td class=\"src\">CyberScoop<\/td>\n<td class=\"dt\">Aug 10, 2026<\/td>\n<\/tr>\n<tr>\n<td>5. <a href=\"https:\/\/www.nextgov.com\/policy\/2026\/08\/gao-official-suggests-addressing-ai-risks-through-existing-legislation\/415410\/\">GAO official suggests addressing AI risks through existing legislation<\/a><\/td>\n<td class=\"src\">Nextgov\/FCW<\/td>\n<td class=\"dt\">Aug 13, 2026<\/td>\n<\/tr>\n<tr>\n<td>6. <a href=\"https:\/\/www.techrepublic.com\/article\/news-white-house-open-ai-model-safety-reviews\/\">White House AI Safety Reviews Could Expand to Open Models<\/a><\/td>\n<td class=\"src\">TechRepublic<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>7. <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/14\/etsi-cyber-resilience-act-standards\/\">17 draft Cyber Resilience Act standards are open for comment<\/a><\/td>\n<td class=\"src\">Help Net Security<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>8. <a href=\"https:\/\/www.scworld.com\/brief\/california-launches-ai-cybersecurity-initiative-amid-growing-threats\">California launches AI cybersecurity initiative amid growing threats<\/a><\/td>\n<td class=\"src\">SC Media<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>The board conversation: tech risk, growth, and the backlog<\/h4>\n<div class=\"cluster-intro\">Four arguments about why security programmes stall that all point at the operating model rather than the security team &mdash; invisible technical debt, a function still framed as the department of no, a backlog nobody owns, and a hiring problem that may be a throughput problem.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>9. <a href=\"https:\/\/www.darkreading.com\/cyber-risk\/what-boards-must-know-tech-risk\">What Boards Need to Know About Tech Risk<\/a><\/td>\n<td class=\"src\">Dark Reading<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>10. <a href=\"https:\/\/www.csoonline.com\/article\/4208202\/how-csos-can-turn-cybersecurity-into-a-business-growth-strategy.html\">How CSOs can turn cybersecurity into a business growth strategy<\/a><\/td>\n<td class=\"src\">CSO Online<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>11. <a href=\"https:\/\/www.csoonline.com\/article\/4209334\/the-cybersecurity-backlog-is-not-a-security-problem.html\">The cybersecurity backlog is not a security problem<\/a><\/td>\n<td class=\"src\">CSO Online<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>12. <a href=\"https:\/\/www.scworld.com\/perspective\/the-cybersecurity-talent-shortage-was-never-real\">The cybersecurity talent shortage was never real<\/a><\/td>\n<td class=\"src\">SC Media<\/td>\n<td class=\"dt\">Aug 12, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>Enforcement and preparedness: what regulators and surveys found<\/h4>\n<div class=\"cluster-intro\">A UK regulator&#8217;s post-mortem on unmonitored patching and ignored malware alerts, and a manufacturing-sector survey showing roughly half the sector without the basics.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>13. <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/ico-reprimands-acro-records-office\/\">ICO Reprimands Criminal Records Office After 2023 Breach<\/a><\/td>\n<td class=\"src\">Infosecurity Magazine<\/td>\n<td class=\"dt\">Aug 13, 2026<\/td>\n<\/tr>\n<tr>\n<td>14. <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/half-uk-manufacturers-cyber\/\">Only Half of UK Manufacturers Have a Cyber Incident Response Plan<\/a><\/td>\n<td class=\"src\">Infosecurity Magazine<\/td>\n<td class=\"dt\">Aug 11, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>Rebuilding the plumbing: post-quantum and the vulnerability database<\/h4>\n<div class=\"cluster-intro\">Two multi-year infrastructure programmes with dates attached &mdash; a cloud provider&#8217;s cryptographic migration schedule, and a federal request for input on what the NVD has to become to survive machine-scale disclosure.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>15. <a href=\"https:\/\/www.securityweek.com\/google-cloud-sets-out-post-quantum-roadmap-with-2029-readiness-goal\/\">Google Cloud Sets Out Post-Quantum Roadmap With 2029 Readiness Goal<\/a><\/td>\n<td class=\"src\">SecurityWeek<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>16. <a href=\"https:\/\/cyberscoop.com\/nist-national-vulnerability-database-ai-overhaul\/\">NIST wants to overhaul its vulnerability database for the AI age<\/a><\/td>\n<td class=\"src\">CyberScoop<\/td>\n<td class=\"dt\">Aug 11, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>Law, researchers, and surveillance transparency<\/h4>\n<div class=\"cluster-intro\">A 1990 statute that still doesn&#8217;t distinguish research from crime, and a transparency commitment on government hacking tools that arrives on a three-year delay.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>17. <a href=\"https:\/\/www.darkreading.com\/application-security\/outdated-cybercrime-laws-security-researchers-risk\">Outdated Cybercrime Laws Put Security Researchers at Risk<\/a><\/td>\n<td class=\"src\">Dark Reading<\/td>\n<td class=\"dt\">Aug 10, 2026<\/td>\n<\/tr>\n<tr>\n<td>18. <a href=\"https:\/\/www.scworld.com\/brief\/u-s-judiciary-to-publicly-disclose-use-of-hacking-tools-in-wiretaps-starting-2029\">US judiciary to publicly disclose use of hacking tools in wiretaps starting 2029<\/a><\/td>\n<td class=\"src\">SC Media<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<\/table>\n<h4>Agentic AI readiness and the infrastructure underneath it<\/h4>\n<div class=\"cluster-intro\">The gap between agentic-AI ambition and operational readiness, the grid buildout that powers it, the argument for fixing fundamentals first, a frontier lab pausing its own model over offensive-cyber capability, and the political backlash forming around all of it.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>19. <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/14\/deloitte-agentic-ai-readiness-gap-report\/\">The hardest part of agentic AI may be rebuilding the business<\/a><\/td>\n<td class=\"src\">Help Net Security<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>20. <a href=\"https:\/\/www.scworld.com\/analysis\/ai-driven-energy-buildout-raises-cybersecurity-supply-chain-risks\">AI-driven energy buildout raises cybersecurity, supply chain risks<\/a><\/td>\n<td class=\"src\">SC Media<\/td>\n<td class=\"dt\">Aug 14, 2026<\/td>\n<\/tr>\n<tr>\n<td>21. <a href=\"https:\/\/www.scworld.com\/perspective\/stop-bracing-for-the-ai-monster-fix-the-boring-stuff-first\">Stop bracing for the AI monster &ndash; fix the boring stuff first<\/a><\/td>\n<td class=\"src\">SC Media<\/td>\n<td class=\"dt\">Aug 11, 2026<\/td>\n<\/tr>\n<tr>\n<td>22. <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/openai-pauses-development-astra\/\">OpenAI Pauses Some Development of Astra Model on Security Concerns<\/a><\/td>\n<td class=\"src\">Infosecurity Magazine<\/td>\n<td class=\"dt\">Aug 11, 2026<\/td>\n<\/tr>\n<tr>\n<td>23. <a href=\"https:\/\/www.transformernews.ai\/p\/the-jan-6-organizer-getting-conservatives-angry-about-ai\">The Jan 6 organizer getting conservatives riled up about AI<\/a><\/td>\n<td class=\"src\">Transformer<\/td>\n<td class=\"dt\">Aug 13, 2026<\/td>\n<\/tr>\n<\/table>\n<h3>Foundational Reading<\/h3>\n<h4>The regulatory baseline behind this week&#8217;s headlines<\/h4>\n<div class=\"cluster-intro\">The context that makes the week legible &mdash; the EU AI Act obligations that came into force on August 2, a diagnosis of why boardroom cyber governance keeps failing, and the GAO&#8217;s count of how much of the US regulatory estate is already duplicated.<\/div>\n<table class=\"index-table\">\n<tr>\n<th>Article<\/th>\n<th>Source<\/th>\n<th>Published<\/th>\n<\/tr>\n<tr>\n<td>24. <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/04\/eu-ai-act-enforcement-ai-models\/\">EU begins enforcing AI Act, putting AI models under the microscope<\/a><\/td>\n<td class=\"src\">Help Net Security<\/td>\n<td class=\"dt\">Aug 4, 2026<\/td>\n<\/tr>\n<tr>\n<td>25. <a href=\"https:\/\/www.cio.com\/article\/4199585\/3-cybersecurity-issues-that-should-keep-every-ceo-awake-at-night.html\">3 cybersecurity issues that should keep every CEO awake at night<\/a><\/td>\n<td class=\"src\">CIO.com<\/td>\n<td class=\"dt\">Jul 24, 2026<\/td>\n<\/tr>\n<tr>\n<td>26. <a href=\"https:\/\/www.cybersecuritydive.com\/news\/cybersecurity-regulations-overlap-harmonization-gao\/826012\/\">GAO report details scope of cybersecurity regulation overlap<\/a><\/td>\n<td class=\"src\">Cybersecurity Dive<\/td>\n<td class=\"dt\">Jul 23, 2026<\/td>\n<\/tr>\n<\/table>\n<p>            <!-- Detailed write-ups --><\/p>\n<h2>Detailed write-ups<\/h2>\n<div class=\"article\">\n<h4>1. The White House authorises private-sector offensive cyber &mdash; and reopens a 300-year-old argument<\/h4>\n<p class=\"meta\">Cybersecurity Dive &middot; CyberScoop &middot; Nextgov\/FCW &middot; August 13&ndash;14, 2026<\/p>\n<p>The memorandum President Trump signed on August 13 directs the Departments of Justice and Homeland Security to build a programme under which vetted private companies may conduct surveillance and sabotage operations against foreign criminal organisations. The guardrails are specific: operations may not kill or seriously injure anyone, may not amount to a use of force or armed attack under international law, must be approved individually by co-executive directors drawn from DOJ and DHS, and must be backed by a $1 million bond that is forfeit if the rules are broken. Companies must notify the government if they accidentally touch a US person or system, and US Cyber Command is in the loop for deconfliction. A classified annex accompanies the public text, and a coordination centre has 60 days to produce operating procedures. Joseph Alm, the DHS assistant secretary for cyber, infrastructure, risk and resilience, is the named official on the DHS side; Amanda Naylor, the NSC&#8217;s cyber policy director, framed the point of the exercise as leveraging private-sector &ldquo;capabilities, speed, and innovation.&rdquo;<\/p>\n<p>The expert reaction split along a line worth understanding, because it is the line your own risk committee will end up arguing over. Supporters &mdash; Ari Redbord of TRM Labs called it &ldquo;a huge step toward empowering the private sector,&rdquo; Huntress CEO Kyle Hanslovan wants a stronger &ldquo;coalition of the willing,&rdquo; Columbia&#8217;s Jason Healey is willing to try it with specific performance criteria &mdash; treat it as an overdue matching of capability to authority against adversaries operating at machine scale. Critics attack the mechanics rather than the ambition. Paul Rosenzweig, formerly DHS deputy assistant secretary for policy, was blunt: &ldquo;This is a bad idea.&rdquo; Erica Lonergan of Columbia questioned how companies get vetted, who sets the goals, and who audits the outcome. Gary Corn, once staff judge advocate at Cyber Command, warned that deconfliction becomes &ldquo;exponentially more challenging&rdquo; once private operators are in the field. Michael Garcia of Monument Advocacy, a former CISA official, went straight to attribution: criminal groups and state proxies are frequently the same people, and a company that gets that wrong may hand a foreign government a pretext. Errata Security&#8217;s Robert Graham asked the durability question &mdash; whether the oversight holds up in year five. Davi Ottenheimer&#8217;s verdict, &ldquo;it&#8217;s an embarrassment to America,&rdquo; is the sharp end of a broader concern that designated targets get no notification, no prevention mechanism and no appeal.<\/p>\n<p>Nextgov\/FCW&#8217;s essay supplies the framing that ties the week together. CISA incident-response trainer Jeff Gray&#8217;s line &mdash; &ldquo;the United States just revived privateering for the digital age&rdquo; &mdash; is not rhetorical flourish; letters of marque are the closest legal precedent, and the parallel carries all of privateering&#8217;s historical problems with it. The essay&#8217;s larger point is that this is continuous with something already under way: Shield AI holding a CENTCOM contract for drone ISR, SpaceX running Starlink for troop connectivity, Anduril maintaining drones alongside special operations forces. Warfare is becoming a procurement-and-engineering activity, and cyber operations are the next line item. One executive quoted anonymously expects &ldquo;every Israeli cyber startup to jump at the chance.&rdquo; For CISOs the practical exposure is threefold: security vendors in your supply chain may soon hold a licence you did not evaluate them against; adversary retaliation calculus changes when private US firms become legitimate-seeming targets; and if your own board asks whether the company should participate, the answer needs to have been thought through before the 60-day guidance lands.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.cybersecuritydive.com\/news\/us-private-companies-gangs-cyberattacks-offensive-operations\/827805\/\">Cybersecurity Dive<\/a> &middot; <a href=\"https:\/\/cyberscoop.com\/private-sector-hacking-presidential-memo-cybersecurity\/\">CyberScoop<\/a> &middot; <a href=\"https:\/\/www.nextgov.com\/cybersecurity\/2026\/08\/warfare-becomes-engineering-era-digital-mercenary-dawns\/415442\/\">Nextgov\/FCW<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>2. Four AI regulatory tracks, one week, no coordination<\/h4>\n<p class=\"meta\">CyberScoop &middot; Nextgov\/FCW &middot; TechRepublic &middot; Help Net Security &middot; SC Media &middot; Transformer &middot; August 4&ndash;14, 2026<\/p>\n<p>The FTC&#8217;s proposed policy statement, released in July and picked apart this week, asks whether ideological bias in AI systems constitutes an unfair and deceptive practice under Section 5 of the FTC Act &mdash; on the theory that consumers expect AI output free of ideological manipulation. The consequential part is not the culture-war framing but the pre-emption: the commission&#8217;s position would override state regimes, naming the Colorado AI Act and its 2027 pre-release bias-audit requirement. Reaction crossed the usual lines. Leah Siskind of the Foundation for Defense of Democracies, formerly a White House digital official, argued the agency&#8217;s job is &ldquo;to police consumer protection violations, not regulating AI systems,&rdquo; and dismissed the underlying dispute as &ldquo;petty squabbles about which AI model is more woke.&rdquo; Cato researchers David Inserra, Jennifer Huddleston and Juan Londo&ntilde;o called the conflation of ideological bias with factual deception &ldquo;an absurd comparison.&rdquo; America First Legal&#8217;s Emily Percival argued the opposite. Anthropic appears extensively in the statement&#8217;s footnotes; xAI barely features.<\/p>\n<p>At the GovCIO AI Forum on August 13, GAO&#8217;s Nick Marinos made a structurally different argument: Congress should amend existing cybersecurity law rather than write AI-specific statutes, because the current approach is reactive. His concrete proposal is to update the Cybersecurity Information Sharing Act of 2015 so its definitions explicitly cover AI-system threat data &mdash; a law whose provisions were due to expire in September and now run on a short extension through December 11, 2026. Michael Daniel of the Cyber Threat Alliance agreed. Meanwhile TechRepublic reported the White House weighing an extension of its voluntary safety-review framework from closed models to open-weight models once they reach frontier capability comparable to Anthropic&#8217;s Mythos class or OpenAI&#8217;s GPT-5.6 &mdash; a move Meta, Microsoft, Palantir and Nvidia have reason to resist, since open weights can be downloaded and modified after release and the compliance burden falls hardest on smaller developers. Brussels moved on schedule rather than in reaction: ETSI&#8217;s TC CYBER-EUSR, chaired by Sandra Feliciano, opened seventeen draft harmonised standards for the Cyber Resilience Act&#8217;s higher-risk product tier &mdash; password managers, anti-virus, smart home assistants, connected toys, wearables &mdash; with comment windows staggered from mid-September to mid-November and a hard compliance date of December 31, 2027. Following a harmonised standard buys presumption of conformity, which makes these drafts the actual specification most product teams will build against. That sits on top of the AI Act obligations already in force since August 2, 2026: chatbots must identify themselves, synthetic content must be labelled and machine-readably marked, GPAI providers must publish training-data summaries and copyright policies, with penalties up to &euro;15 million or 3% of worldwide turnover.<\/p>\n<p>California declined to wait for any of it. Governor Newsom launched an AI cybersecurity initiative run out of the state&#8217;s Cybersecurity Integration Center, mandating an AI cybersecurity officer in every state agency, deploying AI for vulnerability detection, network hardening and incident response, and extending resources to local governments and critical infrastructure operators &mdash; with the governor citing insufficient federal support as the reason. The subtext to all five tracks is political, and Transformer&#8217;s profile of Humans First is the best available read on it: Amy Kremer, a Tea Party veteran who helped secure permits for the January 6 rally, now chairs a conservative anti-AI organisation incubated with Center for AI Safety funding, mobilising Republican voters against AI companies and data centres. &ldquo;Big Tech has all the money,&rdquo; she says. &ldquo;They bought off everybody in Washington DC.&rdquo; The relevance for a CISO is not the politics but the volatility: an enterprise AI governance programme calibrated to the current federal posture is calibrated to something that may not survive its own coalition.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/cyberscoop.com\/ftc-regulating-ai-ideological-bias\/\">CyberScoop (FTC)<\/a> &middot; <a href=\"https:\/\/www.nextgov.com\/policy\/2026\/08\/gao-official-suggests-addressing-ai-risks-through-existing-legislation\/415410\/\">Nextgov\/FCW (GAO)<\/a> &middot; <a href=\"https:\/\/www.techrepublic.com\/article\/news-white-house-open-ai-model-safety-reviews\/\">TechRepublic<\/a> &middot; <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/14\/etsi-cyber-resilience-act-standards\/\">Help Net Security (CRA standards)<\/a> &middot; <a href=\"https:\/\/www.scworld.com\/brief\/california-launches-ai-cybersecurity-initiative-amid-growing-threats\">SC Media (California)<\/a> &middot; <a href=\"https:\/\/www.transformernews.ai\/p\/the-jan-6-organizer-getting-conservatives-angry-about-ai\">Transformer<\/a> &middot; <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/04\/eu-ai-act-enforcement-ai-models\/\">Help Net Security (EU AI Act)<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>3. Boards, backlogs and the throughput argument<\/h4>\n<p class=\"meta\">Dark Reading &middot; CSO Online &middot; SC Media &middot; CIO.com &middot; Cybersecurity Dive &middot; July 23 &ndash; August 14, 2026<\/p>\n<p>Chris Drumgoole, president of global infrastructure services at DXC, makes the cleanest version of the board argument in Dark Reading: technology risk is systematically underpriced because mitigating it produces no visible return. Modernisation and resilience work show up as an absence of failures, never as a line in quarterly earnings, so the risk accumulates silently while teams work around it daily. He names six categories worth a board&#8217;s standing attention &mdash; deferred modernisation, technical debt, thinned AI governance, supply-chain dependency, cloud concentration, and degraded operational resilience &mdash; and cites Accenture&#8217;s 2024 figure that US companies lose close to $2.5 trillion a year to technical debt against roughly $1.5 trillion to resolve it. His two suggested board questions are better than most frameworks: which systems would cause the greatest disruption if they failed, and what keeps our CIO awake at night. The corollary he draws matters as much as the questions: the answers only arrive if the CIO and CISO believe reporting bad news is safe.<\/p>\n<p>CSO Online ran two pieces the same day that fit together. Christopher Squier, CISO and VP for North America at Sodexo, argues the function&#8217;s next evolution is &ldquo;moving from the team that says &lsquo;no&rsquo; to the team that helps organizations say &lsquo;yes&rsquo;&rdquo; &mdash; involved before technology decisions rather than at the final gate, measuring resilience and recovery rather than incidents prevented, and designing controls users barely notice so they stop routing around them. Joshua Copeland, a cybersecurity executive and Tulane professor, supplies the structural counterpart: a large vulnerability backlog is &ldquo;a record of unresolved organizational decisions,&rdquo; not a security-team failure. Security discovers and oversees; infrastructure, cloud and application teams execute; executives break ties. Where that split is not explicit, the backlog grows regardless of how good the security team is. His recommendations &mdash; separately funded time-limited remediation teams, attacking recurring categories rather than individual tickets, and judging security on validation and escalation speed rather than remediation completion &mdash; are unusually actionable, and lean on NIST CSF 2.0, NIST&#8217;s patch-management guidance and CISA&#8217;s KEV catalogue rather than invented metrics.<\/p>\n<p>Anurag Gurtu, co-founder and CEO of Airrived, pushes the same logic one step further in SC Media: the talent shortage was never real, it was a throughput problem. Alert volume grows exponentially with infrastructure while headcount grows linearly, and &ldquo;we cannot out-hire an exponential curve.&rdquo; He points to agentic automation platforms reaching roughly 90% automation of frontline triage in production with mean time to respond down about 80%, and argues the money belongs in automation rather than another hiring pipeline. Read alongside JC Gaillard&#8217;s foundational CIO.com piece &mdash; dashboards measuring activity rather than resilience, governance structures built for an older threat model, and a &ldquo;cybersecurity spiral of failure&rdquo; where more products generate more alerts and reports without fixing anything structural &mdash; and the week&#8217;s board-facing material converges on a single claim: the constraint is the operating model. The GAO&#8217;s regulatory-overlap findings are the external version of the same disease. Thirty-seven agencies have issued 117 cybersecurity rules across nine sectors; 80 of them cover the same ground; roughly 70% contain redundant reporting requirements; financial-services firms may face 15 separate incident-reporting rules, with CIRCIA expected to make the timing conflicts worse.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.darkreading.com\/cyber-risk\/what-boards-must-know-tech-risk\">Dark Reading<\/a> &middot; <a href=\"https:\/\/www.csoonline.com\/article\/4208202\/how-csos-can-turn-cybersecurity-into-a-business-growth-strategy.html\">CSO Online (growth strategy)<\/a> &middot; <a href=\"https:\/\/www.csoonline.com\/article\/4209334\/the-cybersecurity-backlog-is-not-a-security-problem.html\">CSO Online (backlog)<\/a> &middot; <a href=\"https:\/\/www.scworld.com\/perspective\/the-cybersecurity-talent-shortage-was-never-real\">SC Media<\/a> &middot; <a href=\"https:\/\/www.cio.com\/article\/4199585\/3-cybersecurity-issues-that-should-keep-every-ceo-awake-at-night.html\">CIO.com<\/a> &middot; <a href=\"https:\/\/www.cybersecuritydive.com\/news\/cybersecurity-regulations-overlap-harmonization-gao\/826012\/\">Cybersecurity Dive (GAO overlap)<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>4. Two multi-year programmes with dates: post-quantum readiness and the NVD rebuild<\/h4>\n<p class=\"meta\">SecurityWeek &middot; CyberScoop &middot; August 11&ndash;14, 2026<\/p>\n<p>Google Cloud&#8217;s post-quantum roadmap is useful precisely because it commits to dates. Google.com and googleapis.com endpoints already run NIST-standardised ML-KEM key exchange in hybrid mode; application and proxy load balancers support quantum-safe hybrid key exchange for TLS 1.3 on an opt-in basis; Cloud KMS is generally available for NIST-standardised PQC key exchange and digital signatures. By the end of 2027 the target is mitigating store-now-decrypt-later exposure across customer-facing workloads, administrative tooling such as Cloud VPN and Interconnect, and data-transfer services including the BigQuery CLI and Storage Transfer Service. By the end of 2028: quantum-resistant software supply-chain attestations, quantum-safe certificates, Cloud IAM hardening, and the foundational key-management work including quantum-safe key import and hardware-backed protection. General readiness lands in 2029, with Google explicitly noting it expects the work to continue &ldquo;into the 2030s to support broader industry guidance and evolving global standards,&rdquo; and referencing CNSA 2.0 and NIST IR 8547.<\/p>\n<p>The reason to read that as a planning document rather than vendor marketing is the store-now-decrypt-later clock. Data your organisation transmits today with classical key exchange is harvestable now and decryptable later; the 2027 milestone is the one that determines how much of your traffic falls inside that window. A CISO can use this roadmap directly &mdash; it dates the point at which a major hyperscaler expects its own customer-facing exposure closed, which is a defensible reference point for setting an internal crypto-inventory and migration schedule, and a reasonable question to put to every other cloud and SaaS provider in the estate.<\/p>\n<p>NIST&#8217;s request for information on modernising the National Vulnerability Database is the same kind of story from the other direction. The agency&#8217;s own framing concedes that the traditional model &mdash; periodic scanning, static prioritisation, manual remediation &mdash; is becoming obsolete under the combined pressure of rising disclosure volume and complexity, inconsistent data quality, demand for near-real-time enrichment, and AI hacking tools accelerating both discovery and exploitation. The stated aim is &ldquo;a future-ready vulnerability management ecosystem that is continuous, contextual, and automated,&rdquo; with better automation in vulnerability reporting, faster dissemination, transparency and auditability in AI-driven decisions, and an expanded role for AI in automated remediation. No timeline, staffing or funding commitments accompany the RFI, which is the caveat to keep in view: this is a consultation, not a delivery plan. Organisations whose vulnerability management depends on NVD enrichment timeliness should treat the RFI as an invitation to state that dependency in writing, and in the meantime should not assume the enrichment gap closes on its own.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.securityweek.com\/google-cloud-sets-out-post-quantum-roadmap-with-2029-readiness-goal\/\">SecurityWeek<\/a> &middot; <a href=\"https:\/\/cyberscoop.com\/nist-national-vulnerability-database-ai-overhaul\/\">CyberScoop<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>5. Agentic AI outruns the organisations deploying it &mdash; and the grid underneath it<\/h4>\n<p class=\"meta\">Help Net Security &middot; SC Media &middot; Infosecurity Magazine &middot; August 11&ndash;14, 2026<\/p>\n<p>Deloitte&#8217;s readiness research puts a number on a gap most security leaders have been describing anecdotally: only 16% of leaders say their organisation&#8217;s processes are ready for agentic AI, and just 5% call themselves highly prepared. Around half understand how AI agents will change their operating model. Forty-three per cent expect significant to extreme job disruption within 12 to 18 months; 31% expect at least half their business processes redesigned around agents within two years, rising to 74% within four. Seventy-one per cent provide baseline agent-literacy training and 65% are reskilling, but half say they are not investing enough in the workforce changes required. The three named obstacles are the ones that land on a CISO&#8217;s desk: absent unified data infrastructure, insufficient trust and governance frameworks for agents, and integration complexity and cost. Deloitte&#8217;s Laura Shact frames the failure mode as layering agents into existing workflows instead of redesigning the process &mdash; quick wins that cap long-term value, and, from a security standpoint, agents inheriting the permissions and blast radius of workflows never designed for autonomous actors. Governance here means something concrete: which decisions an agent may take alone, when a human must intervene, and who is accountable when it goes wrong.<\/p>\n<p>Klaas Meinke of Hadrian supplies the useful corrective in SC Media. AI mostly accelerates traditional attack methods rather than inventing new ones; Verizon&#8217;s DBIR keeps showing that attackers succeed through known vulnerabilities, stolen credentials and misconfiguration rather than anything exotic. What AI compresses is time &mdash; hours of manual work down to minutes. His list of &ldquo;boring stuff&rdquo; is the actual agentic-AI security programme for most organisations: patching discipline, accurate asset inventory, privileged-access review, credential rotation, closing unnecessary ports, locking down MCP servers, right-sizing agent permissions, and continuous external attack-surface testing. Two of those items &mdash; MCP servers and agent permissions &mdash; did not exist as categories two years ago and now sit in the middle of the hygiene list.<\/p>\n<p>The capability curve is real enough that a frontier lab stopped its own work over it. OpenAI announced on August 7 that it had paused internal testing of Astra after assessing its cybersecurity capability as &ldquo;critical&rdquo; under the company&#8217;s Preparedness Framework &mdash; the model could identify and build working zero-day exploits against hardened systems without human intervention, and devise novel attack strategies against difficult targets. Mitigations include isolated test environments, restricted network access, enhanced encryption and sandboxed execution. ArmorCode&#8217;s Matt Sayar welcomed the restraint; Black Hills Information Security&#8217;s John Strand was less generous about self-policing, noting frontier labs &ldquo;have demonstrated again and again that we cannot simply assume they&#8217;re going to do the right thing&rdquo;; Ridge Security&#8217;s Nick Mo pointed out that open-weight models already carry comparable capability and are already being used maliciously &mdash; which is exactly the argument the White House is having about extending safety reviews to open models. Underneath all of it sits the physical layer: SC Media&#8217;s analysis of the AI-driven energy buildout, drawing on an ICIT panel with Carnegie Mellon&#8217;s Phoebe Vencill, formerly of ONCD, and Maryland&#8217;s David Mussington, formerly a CISA executive assistant director. Digitally native energy systems bring better monitoring and forensics but widen the attack surface and deepen dependence on Chinese-sourced components; Vencill flags vendor firmware access for diagnostics and updates as a compromise path and argues for risk-based sourcing, while Mussington warns that smaller and rural operators face expanding vulnerability without comparable defensive capability.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.helpnetsecurity.com\/2026\/08\/14\/deloitte-agentic-ai-readiness-gap-report\/\">Help Net Security (Deloitte)<\/a> &middot; <a href=\"https:\/\/www.scworld.com\/perspective\/stop-bracing-for-the-ai-monster-fix-the-boring-stuff-first\">SC Media (fundamentals)<\/a> &middot; <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/openai-pauses-development-astra\/\">Infosecurity Magazine (Astra)<\/a> &middot; <a href=\"https:\/\/www.scworld.com\/analysis\/ai-driven-energy-buildout-raises-cybersecurity-supply-chain-risks\">SC Media (energy buildout)<\/a><\/p>\n<\/p><\/div>\n<div class=\"article\">\n<h4>6. Enforcement, exposure and the slow arrival of transparency<\/h4>\n<p class=\"meta\">Infosecurity Magazine &middot; Dark Reading &middot; SC Media &middot; August 10&ndash;14, 2026<\/p>\n<p>The ICO&#8217;s reprimand of the UK Criminal Records Office is a patch-management post-mortem with an unusually clear ownership failure at its centre. Between August 2022 and March 2023, attackers accessed ACRO&#8217;s website and content management system, exposing data on more than 10,920 people &mdash; names, dates of birth, addresses, National Insurance numbers, passport and driving licence details, bank accounts, biometric data and criminal records, including records belonging to International Child Protection Certificate applicants and domestic-violence survivors. The regulator found two failures. First, split patching responsibility with nobody holding the whole: a managed service provider handled operating-system patches while a separate web development supplier handled Kentico CMS patches, and, in the ICO&#8217;s words, &ldquo;ACRO itself did not monitor for required security patches, meaning that there was an absence of oversight for this important security control.&rdquo; Second, Trend Micro malware alerts fired and were never reviewed or acted upon. ACRO has since segmented its network, decommissioned the compromised infrastructure, migrated services and improved monitoring. The outcome was a reprimand rather than a fine, which is the usual public-sector treatment &mdash; and the reason to read this as a control-design lesson rather than a penalty story. Two suppliers each patching part of a stack, with no one accountable for the join, is a pattern that exists in most enterprises.<\/p>\n<p>Make UK&#8217;s <em>Cyber Security in Manufacturing<\/em> report, published August 10, quantifies how common that kind of gap is across a whole sector. Thirty per cent of UK manufacturers experienced a cyber incident in the past year, directly or through their supply chain. Only 51% have a formal incident response plan; 45% have designated senior leadership responsibility for cyber; 23% employ a dedicated CISO. Among those hit, 31% lost production capacity, 31% were late delivering to customers and 23% faced component or material shortages. Nearly a third either have no cyber insurance or are unsure whether their policy covers cyber disruption. If manufacturing sits anywhere in your supply chain, those are the odds on your suppliers&#8217; preparedness.<\/p>\n<p>The legal environment for the people who find these problems, meanwhile, has not kept up. NCC Group&#8217;s director of government affairs, Katharina Sommer, led research on the UK Computer Misuse Act 1990, which criminalises unauthorised access without distinguishing malicious activity from good-faith research &mdash; a serious problem when much legitimate security work necessarily happens without the system owner&#8217;s explicit consent. Only 15 of the 154 countries with cybercrime statutes offer researchers any protection. Portugal introduced a safe harbour for good-faith research in 2025; Argentina, Chile and Panama have partial protections, with Panama notably covering tool developers. Reform was promised in the King&#8217;s Speech in May. Sommer&#8217;s proposed CICIC framework &mdash; regulate conduct not the actor, weigh intent toward improving security, define good faith through timely reporting and no extortion, add institutional oversight, and attach conditions prohibiting DDoS and limiting data retention &mdash; is a workable template. Alongside it, SC Media reported the Administrative Office of the US Courts will begin publishing how often judges authorise network investigative techniques, the hacking tools and spyware used in wiretaps, following pressure from Senator Ron Wyden. The first data appears in the 2028 Wiretap Report, published in 2029, restoring a disclosure absent for nearly two decades &mdash; though limited to real-time interception, not data extracted from remotely accessed devices. Genuine accountability, arriving on a three-year delay.<\/p>\n<p style=\"font-size:13px;color:#6b7280;margin:0;\">Sources: <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/ico-reprimands-acro-records-office\/\">Infosecurity Magazine (ICO\/ACRO)<\/a> &middot; <a href=\"https:\/\/www.infosecurity-magazine.com\/news\/half-uk-manufacturers-cyber\/\">Infosecurity Magazine (Make UK)<\/a> &middot; <a href=\"https:\/\/www.darkreading.com\/application-security\/outdated-cybercrime-laws-security-researchers-risk\">Dark Reading<\/a> &middot; <a href=\"https:\/\/www.scworld.com\/brief\/u-s-judiciary-to-publicly-disclose-use-of-hacking-tools-in-wiretaps-starting-2029\">SC Media (NIT disclosure)<\/a><\/p>\n<\/p><\/div>\n<p>            <!-- Watch list --><\/p>\n<div class=\"watchlist\">\n<h2>Calls to action &amp; watch list<\/h2>\n<ul>\n<li><strong>Decide your position on the offensive-cyber memo before the 60-day guidance lands.<\/strong> Two questions for the next risk committee: does any security vendor in our supply chain intend to seek authorisation, and does the escalation risk from US private-sector operations change anything in our threat model? Both answers are easier to reach now than after the DOJ\/DHS operating procedures publish in mid-October.<\/li>\n<li><strong>Start a crypto inventory against a 2027 date, not a 2029 one.<\/strong> Google Cloud&#8217;s store-now-decrypt-later milestone is end-2027. Use it to set your own: which data flows, if harvested today, would still be sensitive when decryption becomes feasible, and which providers can name a date for closing that exposure?<\/li>\n<li><strong>Comment on the CRA drafts that cover your products.<\/strong> If you build password managers, anti-virus, smart home assistants, connected toys or wearables for the EU market, the seventeen ETSI drafts are the specification you will be audited against by December 31, 2027 &mdash; and the comment windows close between mid-September and mid-November.<\/li>\n<li><strong>Name an owner for every patch boundary.<\/strong> The ICO&#8217;s ACRO finding is the cheapest lesson available this quarter: two suppliers each patching part of a stack, nobody accountable for the join, and malware alerts nobody read. Audit your own split-responsibility boundaries and your alert-review evidence.<\/li>\n<li><strong>Whether the FTC&#8217;s Section 5 theory survives contact with the courts &mdash; and with the states.<\/strong> Watching whether the proposed policy statement is finalised, and whether its pre-emption of the Colorado AI Act&#8217;s 2027 bias-audit requirement is tested before that date arrives.<\/li>\n<li><strong>The December 11 information-sharing deadline.<\/strong> Watching whether Congress uses the extension window to do what GAO&#8217;s Nick Marinos suggested &mdash; update the 2015 law&#8217;s definitions to cover AI-system threat data &mdash; or simply extends it again.<\/li>\n<li><strong>Whether open-weight models get pulled into the safety-review framework.<\/strong> Watching for a concrete capability threshold, and for how Meta, Nvidia and the open-source camp respond if one is set. OpenAI&#8217;s Astra pause is the precedent the argument will be conducted around.<\/li>\n<li><strong>Whether NIST&#8217;s NVD consultation turns into a funded programme.<\/strong> Watching for a timeline, staffing commitment or budget line attached to the RFI. Until one appears, plan on the enrichment gap persisting.<\/li>\n<li><strong>UK Computer Misuse Act reform.<\/strong> Watching whether the reform promised in May produces a good-faith research safe harbour along the lines NCC Group proposed, and whether other jurisdictions follow Portugal&#8217;s 2025 example.<\/li>\n<li><strong>Agentic AI permissions as an audit item.<\/strong> Watching whether MCP server hardening and agent permission right-sizing show up in mainstream audit and assurance programmes, or stay an engineering-team concern while Deloitte&#8217;s 16% readiness figure holds.<\/li>\n<\/ul><\/div>\n<\/td>\n<\/tr>\n<p>        <!-- Footer --><\/p>\n<tr>\n<td class=\"footer\">\n<p class=\"brand\">The CISO Brief<\/p>\n<p>A weekly intelligence bulletin from Security Radar LLC.<br \/>\n            Curated by Paul Davis &middot; <a href=\"mailto:paul.davis@security-radar.com\">paul.davis@security-radar.com<\/a><\/p>\n<p>&copy; 2026 Security Radar LLC. All rights reserved.<\/p>\n<p>Article titles and summaries are excerpted for review and commentary; all linked articles remain the copyright of their respective publishers and authors.<\/p>\n<p>*|LIST:ADDRESS|*<\/p>\n<p><a href=\"*|ARCHIVE|*\">View this email in your browser<\/a> &middot; <a href=\"*|UNSUB|*\">Unsubscribe<\/a><\/p>\n<\/td>\n<\/tr>\n<\/table>\n<\/td>\n<\/tr>\n<\/table>\n","protected":false},"excerpt":{"rendered":"<p>August 16, 2026 &middot; Weekly Edition The CISO Brief A presidential memo hands vetted private companies the authority to hack transnational criminal organisations &mdash; the biggest structural change to who is allowed to conduct offensive cyber operations in a generation. Around it, AI regulation stops being one debate and becomes&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8,12,42],"tags":[],"class_list":["post-5697","post","type-post","status-publish","format-standard","hentry","category-editorial","category-regulations","category-security-industry-news"],"_links":{"self":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/5697","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5697"}],"version-history":[{"count":1,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/5697\/revisions"}],"predecessor-version":[{"id":5721,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/5697\/revisions\/5721"}],"wp:attachment":[{"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5697"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5697"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.cybersecurityinstitute.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5697"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}