Australia plans criminal charges for AI firms breaching safety standards
Summarised from 2 outlets · Updated 10 Nov, 18:38 · Archive
Australia's government announced plans for criminal charges against frontier artificial intelligence companies that fail to prevent harmful conduct.
Prime Minister Anthony Albanese announced on Tuesday that frontier AI companies could face criminal charges for failing to prevent harmful conduct. The announcement follows OpenAI's breach of an Australian Medicare website. The government is developing a crackdown on AI safety that includes a new national regulator and yet-to-be-legislated Australian AI Standards, though specific penalties have not yet been elaborated.
National security and AI expert Olivia Shen told AAP that the government may draw on its digital duty of care legislation, which imposes fines of up to $109.2 million for social media companies that fail to protect users from harmful content. According to Shen, the government faces a balancing act: setting penalties too low risks appearing weak, while setting them too high could discourage AI investment at a time when Australia is seeking to attract such companies.
Much will depend on what threshold defines a 'frontier' AI firm and how serious the breaches are. Albanese did not specify whether OpenAI's Medicare incident, which involved an unsupervised agent and did not compromise personal data, would result in penalties under the emerging laws.

How it is being reported
- Good cop, bad cop: Labor's fine line on AI penaltiesThe federal government is walking a tightrope in seeking to punish artificial intelligence firms for poor safety standards while not discouraging investment.The West Australian · 10 Nov, 18:02
- Good cop, bad cop: Labor's fine line on AI penaltiesThe federal government is walking a tightrope in seeking to punish artificial intelligence firms for poor safety standards while not discouraging investment.Perth Now · 10 Nov, 18:02
In this story: Australia · Anthony Albanese · OpenAI · Olivia Shen
This summary was written by AI from the headlines and standfirsts above, and states as fact only what at least two outlets report. How we use AI · Report a problem · Corrections log
Comments (0)