Five intelligence agencies — CISA in the US, NCSC in the UK, plus counterparts from Canada, Australia, and New Zealand — issued a joint statement this week warning that AI-powered cyberattacks capable of overwhelming organizational defenses are not years away. They're months away.
That's the Five Eyes. All five. Together. That doesn't happen often.
What makes this advisory different from the steady background noise of cybersecurity warnings is what they said about who's most exposed: not governments, not large enterprises. Small and medium organizations that have under-invested in security. Their phrase, not ours. "Sitting ducks."
If you run a 10-to-100-person NGO, public sector team, or small business, this advisory is addressed to you.
What has actually changed
The specific shift the agencies are describing isn't that attackers are smarter. It's that the timeline between vulnerability discovery and active exploitation is collapsing.
Historically, there was a meaningful window. A security researcher finds a flaw, it gets published, organizations patch it, attackers begin exploiting it — that cycle ran on a timeline of weeks. Sometimes months. Security teams could stay ahead of it if they were reasonably diligent.
AI is compressing that window from weeks to days. In some cases, hours.
Frontier AI models can scan public code repositories and CVE disclosures, identify exploitable patterns, and generate working attack code faster than any human team can respond. The advisory authors specifically noted that AI is "fundamentally reshaping the cyber threat landscape" — not as a future projection, but as a present-tense description of what's happening now.
The difference between "months away" and "already happening" is increasingly semantic.
Why small orgs are named specifically
The Five Eyes advisory didn't name small organizations as an afterthought. It's a central point. Large enterprises have security operations centers, dedicated vulnerability management programs, and the budget to run automated patching pipelines. They're still in danger, but they have the infrastructure to respond.
Small and medium organizations — NGOs doing advocacy work, municipal agencies, 30-person consulting firms, regional health clinics — typically don't have a dedicated IT security function at all. The person responsible for your Microsoft 365 tenant is often also responsible for onboarding new hires and fixing the printer. Patching gets deferred. Outdated systems stay in production because replacing them requires budget cycles. Multi-factor authentication gets implemented unevenly.
Those specific gaps — outdated systems, slow patch management, unnecessary internet-connected services, weak access controls, no incident response plan — are exactly what the advisory calls out. Not as general principles. As the specific conditions that make an AI-assisted attack viable against an organization that would otherwise be too small to bother with.
The economics of attacking small orgs is changing. When a sophisticated attack required significant human expertise and time, targeting a 25-person nonprofit wasn't worth the effort. When AI can automate the reconnaissance, exploit development, and initial access at scale — every under-patched WordPress login page and every MFA-exempt admin account becomes worth targeting.
The five things they said to do right now
The advisory came with concrete recommendations. Not a framework. Not a maturity model. A short list of things to act on immediately:
Limit who and what can connect to your systems. This is the least-privilege principle applied practically: your accounting software shouldn't have network access to your HR system. Your staff laptops shouldn't be directly reachable from the internet. Review what's exposed and close what doesn't need to be open.
Harden authentication across the board. If any accounts that matter — admin access, email, financial systems — can be accessed with just a password, that needs to change. Phishing-resistant MFA isn't a stretch goal. The agencies specifically flagged weak access controls as the entry point attackers are using now.
Install security updates fast. The patching window is closing. The advisory is explicit: the time between a vulnerability being disclosed and it being actively exploited by AI-assisted attacks is now measured in days. If your organization runs on a monthly patch cycle, that's too slow for the threat environment being described.
Replace outdated and unsupported technology. If something in your stack is end-of-life — a Windows version, a network appliance, a piece of accounting software — it's not getting patches. That means every vulnerability found in it stays open permanently. The agencies flagged this as a systemic exposure, not a minor housekeeping item.
Build an incident response plan before you need it. The recommendation for tabletop exercises and pre-incident planning is there because organizations that haven't rehearsed a breach response take two to three times longer to contain one. That matters when attackers are moving faster. You don't need a 40-page document — you need a clear answer to "if we wake up to a ransomware notice tomorrow, who calls who and what happens first."
The part that got less coverage
The agencies also recommended that organizations use AI for defensive purposes. That line is easy to skip past, but it's significant.
The reason AI-powered attacks are becoming more dangerous is the same reason AI-powered defense is becoming more effective: these tools can process and correlate signals at a scale and speed that humans can't match manually. Security monitoring, anomaly detection, log analysis — these are exactly the kind of repetitive pattern-matching tasks that AI handles well. A small org that can't afford a 24/7 SOC can now run automated monitoring that catches anomalous login behavior, unusual data access patterns, or unexpected outbound connections without a human watching a dashboard around the clock.
The advisory isn't saying small orgs need enterprise security budgets. It's saying the same AI capability that's creating this threat is available to defend against it.
What to do this week
The gap between "we know this is a problem" and "we've actually done something about it" is where organizations get compromised.
Start with authentication. Pull up your Microsoft 365 or Google Workspace admin panel and check which accounts are running without MFA enabled. If you have admin accounts or shared service accounts without it, that's your highest-priority fix — it can be done today.
Check your patching cadence. When was the last time you applied updates to the systems and software your team uses? If the answer is "we're not sure" or "last month," move that up.
Know what's exposed to the internet. Run a free scan against your public-facing domains using a tool like Shodan or SecurityHeaders.com. Most small organizations are surprised by what's visible.
And read the original advisory. It's available on the CISA website. It's short, it's readable, and it was written specifically to be actionable for organizations without dedicated security staff.
The Five Eyes don't coordinate joint statements lightly. When they use the phrase "sitting ducks" and point specifically at under-resourced small organizations, that's the kind of signal worth acting on before the incidents start showing up in the news.
If you're not sure where your organization actually sits on any of these dimensions, that assessment is straightforward to do and usually turns up a short list of specific fixes. It's the kind of thing we do in a sprint — not a six-month engagement, not a 200-page gap analysis. A working session that ends with your team knowing what to fix and how to fix it.