HLD Pulse · Intelligence advisory · 23 June 2026Action required

Five Eyes: AI cyber attacks on businesses are months away

On 22 June 2026, all five intelligence agencies — Australia, Canada, New Zealand, the United Kingdom, and the United States — issued a rare joint statement warning that frontier AI models will fundamentally transform offensive cyber capabilities for businesses. The timeline is months, not years. HLD Pulse breaks down what the advisory means for your organisation and what to do right now.

Intelligence · AdvisoryAI ThreatsFive Eyes · CISA · NCSCBusiness risk

Executive summary

On 22 June 2026, the cybersecurity agencies of all Five Eyes nations — Australia's ACSC, Canada's Cyber Centre, New Zealand's NCSC-NZ, the UK's NCSC-UK, and the US's CISA and NSA — jointly published an advisory titled “The AI shift in cyber risk: why leaders must act now.” It was signed by the agency heads themselves, which is unusual and signals the severity of the message.

The central warning: frontier AI models — models capable of autonomously scanning infrastructure, identifying vulnerabilities, generating working exploits, and conducting social engineering at scale — are expected to be publicly or commercially available within months. When that happens, the barrier to conducting sophisticated, targeted attacks against businesses collapses.

Critically, the advisory does not treat this as a future theoretical risk. It frames it as an operational planning horizon that demands board-level attention and immediate foundational action. The core message to executives: cyber risk is no longer purely a technical problem — it is a business leadership responsibility, and the window to get the basics right is closing.

Why this advisory is different

Joint Five Eyes advisories are common for specific technical threats — particular malware families, nation-state campaigns, or CVE disclosures. What is rare is a joint advisory signed by the heads of all participating agencies, focused not on a specific incident but on a structural shift in the threat environment.

The advisory does not name a specific attack or actor. It describes a category shift: the arrival of AI models that automate and accelerate every stage of the attack lifecycle simultaneously. It explicitly states that the pace of AI development means “cyber risk assumptions can become outdated in months, not years.”

This means organisations that completed a cyber risk review six months ago may already be operating under an outdated threat model. The agencies are asking boards and executives to treat this as an ongoing leadership obligation, not a one-time audit exercise.

Weaknesses AI will exploit first

The advisory identifies five specific organisational weaknesses that AI-powered attacks are best positioned to exploit — all of them fixable with existing security practice:

Legacy systems

Unsupported infrastructure represents strategic liabilities. AI threat actors will prioritise known unpatched systems as easy, high-value entry points.

Sluggish patching

AI is compressing the window between vulnerability disclosure and active exploitation. Weeks-long patch cycles that were once acceptable are now dangerous.

Unnecessary internet exposure

Systems and services exposed to the internet beyond operational need represent exploitable surface area that AI-driven reconnaissance will find faster than human teams.

Weak identity and access controls

Inadequate authentication, over-provisioned accounts, and absent permission reviews are the low-hanging fruit AI-powered credential attacks will target first.

Absent incident planning

Organisations without rehearsed breach response plans face compounded losses when AI-speed attacks outpace manual investigation and containment timelines.

Five urgent actions for businesses

The Five Eyes advisory condenses its guidance into five foundational actions. These are not novel concepts — they are baseline security hygiene that the agencies say organisations have consistently failed to fully implement. AI-powered attacks will punish those gaps at machine speed.

01

Reduce your attack surface

Audit every system with external connectivity. Challenge whether exposure is operationally necessary. Remove or isolate anything that does not need to be reachable. AI-powered reconnaissance tools will map your perimeter faster than you can monitor it.

02

Accelerate patching — treat it as a business priority

The time between vulnerability disclosure and first exploitation is shrinking. AI can identify affected systems and generate working exploits in hours. Monthly or quarterly patch cycles are no longer defensible for critical systems.

03

Decommission or isolate legacy systems

Unsupported systems cannot be secured. Where decommissioning is not immediately possible, enforce strict network segmentation, remove direct internet paths, and log all access. Legacy systems must be on a documented retirement timeline.

04

Harden identity and access management

Enforce MFA everywhere. Limit privileged access to specific roles with time-bounded grants. Run regular access reviews and remove stale accounts. Implement just-in-time access for administrative functions. AI-assisted credential attacks make blanket privileged access indefensible.

05

Prepare for incidents — assume breach will occur

Test your incident response plan. Train executives and board members on their role in a cyber event. Maintain offline backups. Know your detection and recovery timelines. The advisory is explicit: prevention alone is insufficient. Resilience is the business imperative.

What specifically changes when AI attacks become mainstream

Today, a sophisticated attacker needs skilled personnel, time, and resources to research a target, identify vulnerabilities, craft exploits, and conduct social engineering. That friction is what makes targeted attacks expensive and limits their scale.

Frontier AI models collapse that friction. A model that can autonomously conduct open-source intelligence, parse public vulnerability databases, generate spear-phishing emails in perfect local language, and adapt attack approaches in real-time can do in hours what previously took specialist teams weeks. More critically, it can do this at scale — running simultaneous campaigns against thousands of organisations in parallel.

The advisory specifically flags that AI will lower the skill barrier for attackers significantly. Threat actors who previously could only deploy commodity malware will gain access to capabilities that previously required nation-state resources. The practical implication: every organisation, not just those in historically targeted sectors, becomes a plausible attack target.

For defenders, the corollary is that human-speed security operations — manual alert triage, slow incident escalation, quarterly patching cycles — cannot match what is coming. Automated, AI-assisted defence at the same operational speed as the attacks is not a future aspiration. It is the emerging baseline.

What the advisory demands from leadership

Boards and executives must understand and assess their cyber risk posture — not delegate it entirely to IT.

Cyber leaders must be empowered with authority and resources proportional to the threat, not constrained by legacy budget structures.

Cyber risk must be integrated into core business planning, M&A decisions, product development, and vendor selection.

Organisations must plan for incidents occurring, not just attempt to prevent them — resilience and recovery are as important as protection.

Risk assumptions must be reviewed regularly; a six-month-old threat model may already be materially outdated.

Advisory timeline

  1. Five Eyes agencies (ACSC, CCCS, NCSC-NZ, NCSC-UK, CISA, NSA) jointly publish "The AI shift in cyber risk: why leaders must act now" — a rare unified statement signed by the heads of all six agencies.

  2. Advisory explicitly warns that frontier AI models capable of transforming offensive cyber capabilities are "months away" from public availability — framing this as an imminent operational risk, not a theoretical future concern.

  3. Agencies urge immediate action from boards and executives, framing cyber risk as a core business leadership issue rather than a technical IT matter.

Sources

Primary source: the official Five Eyes joint advisory published on 22 June 2026 by NCSC-UK, CISA, ACSC, CCCS, and NCSC-NZ. HLD Pulse is an interpretive intelligence service; this briefing is based on publicly available advisory content and independent press corroboration. HLD did not participate in the advisory's preparation.

HLD Pulse

Need a tailored intelligence briefing?

HLD Pulse can map this Five Eyes advisory to your specific technology stack, risk profile, and current security posture — with actionable priorities for your board and operations teams.

View HLD PulseRequest briefing support
HLD Sentinel

AI-speed defence for AI-speed threats

The Five Eyes advisory describes a threat environment that requires autonomous, machine-speed security response. HLD Sentinel delivers exactly that — continuous protection that matches the pace of the attacks coming your way.

Explore Sentinel