Publication,

AI Guides | AI & Liability

AU | EN
Current site :    AU   |   EN
Australia
Singapore

Who is liable when AI fails?

In the 1980s, the Therac-25 radiation system delivered at least 6 fatal or near-fatal doses of radiation to cancer patients. This was due, in part, to a glitch in the computer coding. There were also issues with the design of the system. Further problems were introduced when the system was upgraded at various hospitals. The
correct apportionment of liability is still debated today. So far, in Australia, the legislative focus has been on regulating liability for limited and specific forms of AI, such as autonomous cars and drones. There is no legislation dealing with liability for damage caused by AI generally.

Our experts look at regulating liability for different forms of AI in Australia.

Download
AI Guides: AI & Liability
Our experts look at regulating liability for different forms of AI in Australia.

Download

844KB, 1 Pages

Latest Thinking
Insight
The Department of Home Affairs (Department) has released the Streamlining and Modernising the Security of Critical Infrastructure Act 2018 Consultation Paper (Consultation Paper) here, proposing a second tranche of reforms to the Security of Critical Infrastructure Act 2018 (Cth) (SOCI Act).

15 July 2026

Insight
From 1 July 2026, in New South Wales it is now mandatory for persons conducting a business or undertaking to comply with Approved Safety Codes of Practice.

14 July 2026

Insight
In this edition of the Rail Round-up, we summarise the latest developments across the rail industry, with a particular focus on the recent budget announcements and significant shifts in national rail policy.

10 July 2026