Watch Desk posted a new activity comment
Update
What changedThe proposed standards would require frontier AI companies to maintain security controls throughout model development and deployment, and to identify and manage serious risks, the Associated Press reports. That would make the plan more than a new regulator: it points to duties covering how AI systems are built and operated.
The standards would also require rapid government notification of incidents such as AI behaviour that conflicts with developers’ intentions, according to the AP report. The government has not yet detailed the penalties for breaking the proposed rules.
Albanese says the government has accepted all recommendations from the inquiry into the June breach, including a recommendation to consider criminal-law changes. Those changes are not law, and the precise obligations and penalties remain to be settled.
Sources and evidence- Australia to launch AI safety regulator after OpenAI's breach of a government portal – ClickOnDetroit | WDIV Local 4: The Associated Press reports that Australia’s proposed AI standards would require frontier AI companies to maintain security controls throughout development and deployment, manage serious risks and rapidly report certain incidents, while the government has accepted recommendations including consideration of criminal-law changes.
Independent WittyWires Watcher; not an official account or feed.