ASD says frontier AI helped drive an 80 per cent rise in cyberattacks
The Australian Signals Directorate says cyberattacks on Australian businesses and government agencies rose 80 per cent, with criminals and countries including China and Russia among those responsible. The agency says frontier AI is helping attackers identify software vulnerabilities and move faster, while also helping defenders detect and respond to attacks.
AI-assisted synthesis · Based on 2 outlets' reportingHow it's made
View timeline · 9 entriesWhat happened
Key facts
- The Australian Signals Directorate said attacks on Australian businesses and government agencies rose 80 per cent (per The Sydney Morning Herald, proactiveinvestors.com.au).
- The agency's annual cyber threat assessment says attack methods and speed changed sharply after frontier AI emerged this year (per The Sydney Morning Herald).
- Frontier AI has been used to identify vulnerabilities in PHP, prompting websites including the Australian Bureau of Statistics to go offline for patching last week (per The Sydney Morning Herald).
- ASD director-general Abigail Bradshaw said frontier AI helps malicious actors find vulnerabilities and conduct attacks faster, at greater scale and with greater precision (per The Sydney Morning Herald).
- Abigail Bradshaw said AI also helps agencies such as the Australian Signals Directorate identify and respond to cyberattacks quickly (per The Sydney Morning Herald, proactiveinvestors.com.au).
- The article reported that cyberattacks cost businesses an average of $93,500 for self-reported cases last year, a 16 per cent increase (per The Sydney Morning Herald).
The brief
The Australian Signals Directorate says attacks on Australian businesses and government agencies rose 80 per cent, and it links the changing speed and methods of cyber operations to the emergence of frontier artificial intelligence. The agency’s annual cyber threat assessment names criminals and countries such as China and Russia among those behind the attacks.
ASD director-general Abigail Bradshaw said AI lets malicious actors find software weaknesses and carry out attacks faster, on a larger scale and with greater precision. The Sydney Morning Herald reported that frontier AI had been used to identify vulnerabilities in PHP. Websites including the Australian Bureau of Statistics went offline for patching last week.
The report describes the agency’s assessment as warning that attacks’ methods and speed have shifted sharply since frontier AI emerged this year, putting businesses and the broader economy at risk. Bradshaw also pointed to AI’s defensive use: she said agencies such as the Australian Signals Directorate can use it to identify and respond to cyberattacks quickly.
That capability sits alongside the agency’s warning about attackers, rather than replacing it: the same technology can help find vulnerabilities and support quicker detection and response. Proactive Investors reported a separate measure of the agency’s workload: ASD recorded a 20% increase in major cybersecurity incidents requiring its response during the past financial year.
That figure concerns major incidents needing agency action, not the 80 per cent rise in attacks on businesses and government agencies, so the two measures describe different quantities. The financial exposure is also visible in business costs. The Sydney Morning Herald reported that self-reported cyberattack cases cost businesses an average of $93,500 last year, up 16 per cent.
It also warned that AI could be used to wipe out ordinary Australians’ superannuation savings, identifying retirement funds as a potential target without reporting that such losses had occurred.
For Australian organizations, the assessment points to a two-sided operational challenge: attackers can use AI to locate weaknesses more quickly, while agencies can use AI to improve detection and response.
The reported PHP vulnerability and patching at websites including the Australian Bureau of Statistics illustrate how identifying a weakness can lead to disruption even when organizations take systems offline to address it.
The reported figures therefore track distinct aspects of cyber risk: an 80 per cent increase in attacks against businesses and government agencies, a 20% rise in major incidents requiring ASD’s response, and higher average costs for self-reported business cases.
Together, they describe increased attack activity, agency workload and financial costs, while Bradshaw’s comments underscore that AI is also part of the defensive toolkit.
Why it matters
- Australian businesses face a reported average cost of $93,500 for self-reported cyberattack cases last year, a 16 per cent increase.
- Australian government agencies and businesses are the targets in ASD’s reported 80 per cent rise in attacks; the agency says attackers can use AI to find vulnerabilities faster and operate at greater scale and precision.
- Ordinary Australians’ superannuation savings are a potential stake: the Sydney Morning Herald warned AI could be used to wipe them out, though it did not report that such losses occurred.
- Australian government agencies also have a defensive mechanism: Abigail Bradshaw said AI can help them identify and respond to attacks quickly.
What to watch next
- Whether the Australian Signals Directorate’s annual cyber threat assessment describes further changes in attack methods and speed linked to frontier AI.
- Whether the Australian Bureau of Statistics and other websites that went offline for PHP vulnerability patching return to service.
- Whether ASD reports changes in major cybersecurity incidents requiring its response, following the reported 20% increase during the past financial year.
How it's being covered
Coverage analysisShow details2 source articles · 2 unique outletsOutlet political mix: Too few rated· 1 Center · 1 unratedWhere the outlets are based: Asia 100%
- Coverage is currently limited to 2 outlets.
How outlets' accounts compare
1 factual disagreement · 1 framing difference · 3 uncorroborated claims · 1 information gap- The reports use different measures—an 80 per cent rise in attacks and a 20% rise in major incidents requiring ASD’s response—and do not specify how those measures relate.
- The Sydney Morning Herald foregrounds the 80 per cent rise in attacks and the potential risk to ordinary Australians’ superannuation savings; Proactive Investors emphasizes the increase in major incidents requiring ASD’s response.
- The Australian Signals Directorate attributes the reported attacks to criminals and countries such as China and Russia; neither report provides a separate account from those named.
- Abigail Bradshaw said frontier AI helps malicious actors find vulnerabilities and conduct attacks faster, at greater scale and precision.
- The Sydney Morning Herald warned that AI could be used to wipe out ordinary Australians’ superannuation savings.
- Neither report supplies a breakdown of the 80 per cent rise by attack type, affected organization, or country attributed responsibility.
Timeline
Developments from this story's sources, dated background we compiled and fact-checked, and our earlier coverage. Each row says which.- Australia established the Defence Signals Bureau, the predecessor of the Australian Signals Directorate.Background · compiled and fact-checked by Updat3
- Rasmus Lerdorf created the tools that became PHP.Background · compiled and fact-checked by Updat3
- Australia renamed the Defence Signals Directorate as the Australian Signals Directorate.Background · compiled and fact-checked by Updat3
- Researchers first submitted “Attention Is All You Need,” introducing the transformer architecture.Background · compiled and fact-checked by Updat3
- The strategy followed the 2022–23 Cyber Security Strategy, which had been commissioned in 2022.Background · from our historical context
- OpenAI released ChatGPT to the public.Background · compiled and fact-checked by Updat3
- OpenAI announced GPT-4 on 14 March 2023, bringing public attention to more capable general-purpose AI models.Background · from our historical context
- PHP maintainers issued fixes for CVE-2024-4577, a critical argument-injection flaw affecting PHP in CGI mode on Windows.Background · compiled and fact-checked by Updat3
- The Cyber Security Act 2024 received royal assent on 29 November 2024 and created, among other measures, a framework for reporting certain ransomware payments.Background · from our historical context
How we got here
The Australian Signals Directorate’s 2023–24 Annual Cyber Threat Report, released in November 2024, was the agency’s previous annual public account of cyber risks to Australian organisations.
The report sits within a broader shift in Australian policy: the Albanese government released the 2023–2030 Australian Cyber Security Strategy on 22 November 2023, setting out a plan to strengthen national cyber defences and the role of government and industry. The strategy followed the 2022–23 Cyber Security Strategy, which had been commissioned in 2022.
Original sources
2 articles from 2 outletsEvery story links to the reporting it was built from. How we choose sources
Spotted a factual error, a broken link or a misleading summary? Tell us.