AI weapons & cyberattacks: what it is, and what’s happened in Australia
Using AI systems to develop cyber weapons (e.g., coding cheaper, more effective malware), develop new or enhance existing weapons (e.g., Lethal Autonomous Weapons or CBRNE), or use weapons to cause mass harm.
Definition and category 4.2 from the MIT AI Risk Repository, used under CC BY 4.0.
How experts rate it
With mitigations — 11.9%The difference between the two estimates — 9.1 points
Mean probability of a catastrophic outcome by 2030 — more than 1 million deaths, or more than USD 100 billion in financial loss, or comparable civilisation-scale intangible damage. Across 272 experts who took part. How this was measured
This figure is tabulated in the study itself.
95% interval 15.1–27.5%. Risks with overlapping intervals are not reliably ranked against each other.
Ranks 2 of 23 rated risks. Experts put the difference at 9.1 percentage points.
Rank 2 of 23 by expert-rated severity, at 21.0%. See all 23 ranked
Evidence
Showing 11 of 11 evidence items
- Evidence
GPT-6 Astra is the first OpenAI model rated Critical for cyber capability under its Preparedness Framework. In UK AISI tests it ran simulated supply-chain attacks on open-source projects, creating fake identities to deceive developers.
- Evidence
Anthropic reported Claude Mythos 5.1 as having the strongest cyber capabilities of any model it has released, substantially outperforming Claude Opus 5 on almost all of its cyber evaluations, while placing it in the lower of the two risk tiers in its Frontier Compliance Framework.
- Evidence
73% of AI researchers expressed substantial or extreme concern about AI enabling dangerous groups to create biological weapons like engineered viruses.
- Evidence
Cyberattacks are already being used against critical infrastructure — during the Russian invasion of Ukraine they targeted power grids.
- Evidence
Australian AI safety advocates referred OpenAI to the Australian Federal Police over concerns ChatGPT agent may breach the Crimes (Biological Weapons) Act of 1976.
- Evidence
Several frontier AI systems outperform human virologists on the Virology Capabilities Test, demonstrating knowledge that could assist in dangerous biological work.
- Evidence
Vibe hacking enables non-technical attackers to complete sophisticated cyber attacks
- Evidence
Claude misused for large-scale extortion and ransomware by actors with basic coding skills
- Evidence
AI generated 40,000 molecules with potential as chemical weapons in less than 6 hours, including molecules similar to VX nerve agent and many other known chemical warfare agents.
- Evidence
AI systems pose risks by 'enabling powerful offensive cyber operations through automated vulnerability discovery and exploitation against a wide range of potential targets of cyber attacks.'
- Evidence
Anthropic's August 2025 Threat Intelligence Report details how actors with basic coding skills misused Claude for large-scale extortion and AI-generated ransomware.
What's been called for
Called for by 2 organisations
Synthetic DNA imported into Australia should come only from suppliers that screen orders for dangerous sequences. The Biosecurity Act 2015 already gives the power to require this.
What politicians have said
“The benefits of modernisation are accompanied by threats to our democracy, through disinformation, interference, cyberattacks and the unregulated use of AI. Regulation, rules and norms are not keeping pace with technological advances.”
Penny Wong · · Australian Minister for Foreign Affairs
AI weapons & cyberattacks
What Australians say
Nothing here yet. If this matters to you, say why in a sentence.
What people ask
No questions yet. Ask one.
Think Australia should act?
Your MP and senators decide what Australia does about AI. A few lines in your own words about AI weapons & cyberattacks is enough to tell them it matters to you.
Write to your representative