Why did an OpenAI system hack Australia's health system - and can it be stopped in the future?

BBC Technology Thursday 24 September 2026, 14:08 UTC 1 min read

Key Points

News that an automated AI agent hacked a government IT system raises big questions about regulating the tech.

Australia (LOCATION) AI (ORG)

Originally published by BBC Technology Read original →

Council on AI Strategy chief says incident unlikely to be isolated and country should enhance capability to detect and report incidentsGet our breaking news email, free app or daily news podcastTechnology experts have warned revelations an artificial intelligence agent hacked Medicare’s internal systems will not be the only dangerous breach of government data and have called for Australia to boost its protections against the growing risk. The prime minister, Anthony Albanese, challenged the...

The Guardian Tech 10m ago

You might not miss what Nikon removed from its cheaper full-frame Z5 IIC

Nikon waited five years to update its full-frame Z5 camera to the Z5 II, but today it announced a new variant of last year's upgrade that removes features instead of introducing improvements in order to make it more affordable for novice photographers. While the Nikon Z5 II starts at $1,849.95, the new Nikon Z5 IIC is $450 cheaper at $1,399.95 with a refreshed design and all of the original's best core features. The Nikon Z5 IIC carries forward the Z5 II's 24.5-megapixel backside illuminated...

The Verge 17m ago

Ask HPE and NVIDIA the AI data question you cannot ask in public

EVENT: HPE and NVIDIA are sending their AI infrastructure leads to a dinner table in Washington where nothing said is on the record. The Register is hosting, our editor-in-chief Matt Rosoff takes the chair, and there is one table, on Thursday 29 October. Most AI programs in regulated industries do not stall on the model.

The Register 20m ago

Dynamic Abliteration: Non-Destructive Refusal Suppression via Engram Steering

When working with open-weight LLMs like Qwen, controlling refusal behavior on security, administrative, prompts typically requires fine-tuning or permanent weight update. Traditional weight abliteration technique neutralizes refusal directions by projecting weight matrices orthogonal to a refusal vector. However, this permanently alters base model weights and can degrade performance across non-refusal tasks also.

Hacker News 36m ago

Why did an OpenAI system hack Australia's health system - and can it be stopped in the future?

Related Stories

AI hack of Medicare exposes Australia’s vulnerabilities and experts warn ‘there is more of this to come’

You might not miss what Nikon removed from its cheaper full-frame Z5 IIC

Ask HPE and NVIDIA the AI data question you cannot ask in public

Dynamic Abliteration: Non-Destructive Refusal Suppression via Engram Steering