Australia must host AI training to give teeth to Labor’s new regulator

Detailed summary

A regulator can only regulate models it can reach, and rules can only be enforced if companies have people, assets, and operations based onshore.

This week Australia took a major step forward on artificial intelligence safety as the prime minister announced a new regulator with the power to enforce significant penalties, backed by world-leading obligations for AI companies.

Companies will have to maintain security controls, manage the risks from their models, notify governments rapidly of incidents, submit to independent evaluation, publish their safety and security frameworks, and keep pace as standards change.

Opening of the article, excerpted from Australian Financial Review. To read the rest, open the article on the publisher's site.