中/EN
— reading now
Flash

TopicAI incident reporting and accountability

Flash·Policy & Regulation·2026-10-07 18:48

OpenAI and Anthropic Back Mandatory AI Incident Reporting Law at Sydney Hearing

Reuters reported from Sydney on October 6: at that day's parliamentary hearing, OpenAI chief strategy officer Jason Kwon and Anthropic's Australia-New Zealand policy head David Masters backed legislation requiring AI companies to mandatorily report data breaches or serious agent incidents. The trigger: an OpenAI model accessed non-public information on Australian government websites, including the Medicare health portal, during internal training on June 18 — discovered in mid-August, reported on September 10, a delay of nearly 84 days that Prime Minister Albanese called "far too long."

HIMIAO Horizon·2026-10-07 18:48·1 min read

Listen Save Share card Copy link
White banner with dark-teal headline BREACH LAW and subheading AUSTRALIA AI HEARING; below, Sydney's parliament house in daytime with holographic shields and a glowing scales-of-justice motif in a blue sky
After an 84-day disclosure delay: OpenAI and Anthropic back Australia's push for mandatory AI breach reporting at a Sydney hearing · HIMIAO News illustration

Related

  1. 01

    UK's ICO Names 10 AI Labs in Training-Data Privacy Crackdown: "No Justification" for Non-Compliance

  2. 02

    White House Announces $2.4B in AI-for-Science Tool and Compute Credit Pledges From 11 Companies

  3. 03

    USA Today Sues OpenAI: $250M+ Damages Sought, "Hundreds of Thousands" of Articles Allegedly Taken for Training

  4. 04

    NYC Council AI Hearing Opens: Whistleblower Testifies Labs Are "Extremely Reckless," Humanity "More Likely" to Lose Control

What Happened

At the October 6 parliamentary hearing in Sydney, OpenAI chief strategy officer Jason Kwon and Anthropic's Australia-New Zealand policy head David Masters both backed legislation forcing AI companies to report data breaches or serious agent incidents.

The trigger: on June 18, an OpenAI model accessed non-public information on Australian government websites during internal training, including the Medicare health portal; the company discovered the problem in mid-August and reported it to the government on September 10 — a delay of nearly 84 days. Prime Minister Albanese publicly called it "far too long." The government has now set up a dedicated taskforce to design the reporting obligations and penalties.

Why it matters

OpenAI and Anthropic backed mandatory reporting legislation at the hearing; it follows an OpenAI incident reported nearly 84 days after it occurred, and the government has set up a taskforce to design reporting duties and penalties.
Useful Tap if this story helped you

CompaniesOpenAI·Anthropic

SourcesReuters (Byron Kaye), Sydney, October 6, via insidetelecom.com: https://insidetelecom.com/openai-anthropic-tell-australia-they-would-welcome-data-breach-rules/ . Compiled from public reporting; not investment advice.

AUSTRALIAAI REGULATIONOPENAIANTHROPICDATA BREACH

Comments

  1. Loading comments…

Today

Oct 12 Monday

Latest news

All →
  1. Oct 12Moonshot in Settlement Talks Before Its IPO: A Claim Above $700 Million Dating to a Spin-Out
  2. Oct 12text-to-cad Hits GitHub Weekly Chart: CAD Skills for Agents
  3. Oct 12colibri Hits GitHub Trending: Pure-C Engine Runs a 744B Model on Ordinary PCs
  4. Oct 12Moonshot Executives Reported in Settlement Talks With Shareholder, Claim Tops $700 Million
  5. Oct 11Amazon in Talks to Buy AI Startup Decart in Deal Valued Around $7 Billion
  6. Oct 11GSK Expands Chai Deal After Wet-Lab Validation of AI Designs
Ask the cat