The Australian government is developing a new law to establish a dedicated watchdog tasked with monitoring the security of artificial intelligence systems. Prime Minister Anthony Albanese said the regulator will oversee mandatory cybersecurity measures and will be empowered to issue “significant penalties” against firms that fail to comply with safety rules.
The new framework requires AI companies to notify the government immediately if their models exhibit misaligned behavior or experience cyber incidents. The mandate, which includes requiring companies to submit their models for safety evaluations, follows a June breach of a government website by an OpenAI agent.
