Safety & Security

Two-thirds of Australian firms hit by AI-linked cyber threats: ESET

ESET's survey of 400 Australian security leaders finds 66% hit by AI-linked threats, 92% deploying AI, and confidence outrunning detection performance across the board.

Australian firms face wave of AI cyber threats, ESET says
Australian firms face wave of AI cyber threats, ESET saysVISHNU_KV / pixabay
By James Calloway6 min read

Why it matters

  • 66% of Australian businesses faced at least one AI-related cyber threat in the past year, per ESET's survey of 400 security decision-makers.
  • 92% of surveyed organisations are using or piloting AI in at least one business function.
  • 72% experienced at least one major cybersecurity incident in the past year; 35% of those had three or more.
  • 83% express confidence in cyber resilience despite 53% citing delayed detection as their top containment challenge.
  • 95% report challenges obtaining or maintaining cyber insurance, with rising premiums the top barrier at 47%.

Sixty-six percent of Australian businesses experienced at least one AI-related cyber threat in the past year, according to a survey of 400 cybersecurity decision-makers and influencers across Australia conducted by ESET, the Slovakia-headquartered security vendor. The same research, published 16 November 2026, found that 92% of organisations are already using or piloting AI in at least one business function — a gap between adoption speed and defensive readiness that sits at the centre of the report.

The numbers matter beyond Australia. They quantify, for one of the first times in survey form, how quickly AI has become both an operational staple and an attack surface, with employee misuse, data leakage and synthetic-media fraud all appearing in the results. What ESET does not say is how severe those AI-linked incidents were, only that they occurred — a caveat worth holding against every percentage that follows.

Where is AI showing up — and where is it leaking?

Adoption is broad rather than deep. The surveyed organisations reported using AI in fraud detection, customer service, content generation, document processing, threat detection and software development. No single function dominated.

The risks split between internal and external actors:

  • 34% cited employee misuse of generative AI.
  • 33% reported data leakage through AI platforms.
  • 30% faced AI-generated phishing or impersonation attacks.
  • 29% dealt with AI-enabled deepfake or voice-cloning attacks.

Roughly a third of firms, in other words, are watching their own staff feed corporate data into tools they do not control, while nearly a third are already seeing attackers weaponise the same class of models against them. Safeguards exist, but coverage is partial: 89% of respondents had at least one policy or safeguard in place or planned, while 11% had none at all.

What are firms actually doing about it?

The most common controls are monitoring and restriction, not prohibition. ESET reported that:

  • 49% monitor access to AI tools and their outputs.
  • 48% have internal responsible-use guidelines.
  • 47% restrict uploading sensitive data to AI platforms.

None of these figures reaches a majority of firms, which means the most prevalent safeguard — access monitoring — still leaves over half the sample without it. Parvinder Walia, President of the APAC region at ESET, framed the stakes in operational terms: "AI is becoming deeply embedded in how businesses operate. As we give it access to more data and greater influence over decisions, organisations must ensure the right oversight and safeguards are in place. Trusting AI also means knowing how and where it is being used."

The takeaway in plain words: most firms have some rule on paper, but enforcement and visibility remain the minority position.

How bad is the broader incident picture?

The pressure extends well past AI-specific threats. Some 72% of organisations reported at least one major cybersecurity incident in the past year, and 35% of those hit suffered three or more. Cloud environment breaches and business email compromise (fraudulent emails impersonating executives or suppliers to trigger payments) were the most common incident types, followed by credential theft or abuse and insider threats.

Detection speed looks respectable on paper. ESET found 80% of affected organisations detected and responded to their last significant incident within 24 hours. The containment picture is less flattering:

  • 53% named delayed detection as their leading containment challenge.
  • 40% cited limited internal resources.
  • 38% struggled with coordination across teams.
  • 36% lacked visibility across environments.

Walia's verdict on the speed question was blunt: "In cybersecurity, speed changes the outcome. A threat left undetected for hours can quickly become a business-wide incident. Organisations need continuous visibility, rapid response and access to deep expertise to contain threats early and protect business continuity."

Notably, 53% report delayed detection as a problem even though 80% claim sub-24-hour response — two figures that sit uneasily together, and ESET does not reconcile them.

Is confidence matched by capability?

It is not, by ESET's own reading. The report found 83% of respondents moderately or very confident in their organisation's ability to withstand and recover from a cyberattack. Yet the same sample reported recurring incidents, delayed detections and resource gaps. ESET itself concluded that confidence "does not always align" with response performance — a measured way of saying the industry's self-assessment runs ahead of its results.

What is causing incidents in the first place?

No single cause dominates. Among affected organisations, ESET reported:

  • 37% weak, stolen or compromised passwords.
  • 34% phishing or social engineering attacks.
  • 34% user actions or human error.
  • 30% unpatched vulnerabilities or outdated software.
  • 30% limited security resources or skills shortages.

The human layer remains the dominant failure point, and training has not fixed it. Fully 92% of organisations deliver cybersecurity training at least annually — yet employee awareness and training was still the top cybersecurity challenge, cited by 50% of respondents. The barriers to better training were concrete: 49% named time constraints across teams, 44% pointed to low user engagement, and 38% said the materials themselves were not engaging enough.

Skills shortages compound the problem. A shortage of skilled cybersecurity professionals was cited by 41%, and 40% flagged the pace of change in the threat environment as a prominent concern.

Walia argued the answer is persistence, not technology alone: "Phishing and social engineering continue to demonstrate why the human layer of cybersecurity cannot be overlooked. As AI makes malicious communications more convincing and personalised, organisations need to continuously build employee awareness through relevant training and realistic simulations. Technology can stop many threats, but a well-prepared employee can help stop the one that gets through."

Nearly universal training plus 50% still calling awareness their top challenge is the report's quiet indictment of compliance-style security education.

Where is the money going?

Spending is shifting toward outsourced detection and insurance. ESET found 69% of organisations using or planning to adopt managed detection and response (security monitoring run by an external provider). Cyber insurance coverage stands at 48%, with another 33% planning to obtain cover.

Getting and keeping that insurance is the harder part. Some 95% of organisations reported challenges obtaining or maintaining cyber insurance:

  • 47% cited rising premiums.
  • 42% faced stricter insurer security requirements.
  • 42% had limited internal budgets.

The financial sting of incidents remains real: among organisations affected by incidents, 14% reported significant financial loss. ESET did not define the threshold for "significant," so that figure should be read as directional rather than precise.

Walia closed on the business-continuity argument: "Cybersecurity is now essential to keeping businesses running. As AI and connected technologies grow, organisations need to understand where their risks are, react quickly when something goes wrong, and adopt new technology with confidence."

The report's own data suggests that reaction speed is exactly where confidence and capability diverge — and as AI adoption approaches universality in this sample, the 66% already hit by AI-linked threats is a figure insurers, boards and CISOs will be quoting through 2027.

Source: itbrief.com.au

Share this article:

More from James Calloway

James Calloway

Show full bio

News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.

606 articles

Related articles

  1. AISA warns: Fix cyber basics before buying AI security
  2. BlackFog survey: 47% report sensitive data exposure via AI
  3. One in three AI vendors carries a critical unpatched flaw
  4. INTERPOL: AI Supercharges Existing Cybercrime, Not a New Threat
  5. Australia Warns AI Will Industrialise Scams as Losses Hit $2.18 Billion

« Previous articleNext article »